TigerHawk

[Editor's note: The following is a guest post from an investment professional who wants to be referred to as "Hodge." Hodge is an old friend of mine and a regular TigerHawk reader, and smarter than most really smart people. The post came in last night, but since I was only able to get it up late morning today it does not contemplate today's news. All comments, including from our esteemed co-bloggers (who of whom are also "investment professionals," are more than welcome.]

Where should public money be spent to ensure the United States has a functioning banking system? On deposit insurance, full stop. Buying subordinated bank capital makes the government a borrower from the bank’s senior lenders, which is actually damaging to the capital markets. Guaranteeing toxic bank assets is the same thing as buying subordinated bank capital. Most existing money center banks are and should be thought of as "bad" banks. Government money should be spent on deposit insurance to attract new subordinated risk capital to make new banks to make new loans.

Examine a collateralized debt obligation

Think of a bank as a leveraged financing vehicle, like a big collateralized debt obligation (a bond backed by securitized loans, or "CDO"). A CDO buys a portfolio of credit risk assets – corporate or consumer loans, unsecured debt, residential or commercial mortgage-backed securities – and funds them by issuing layers of debt (“tranches”). The most junior debt tranches get more of the available interest; they are also first in line to take losses if the assets go bad. The senior tranches get less yield but are protected from some portfolio losses.

Subordinated tranche lenders consume leverage – they borrow money from the senior tranches. Senior tranche lenders provide leverage. The senior tranche lenders want asset pricing to be stable — i.e., they want credit spreads not to be volatile. They will also accept the lowest yields if they believe they can move in and out at prices close to par – so they would like to be able to reprice their own spreads and subordination levels as often as possible. The subordinated tranche lenders want the senior tranche lenders to provide permanent, fixed-cost financing if at all possible. If they have that, subordinated tranche lenders can actually use asset spread volatility to their advantage.

Senior and subordinated holders have divergent interests when it comes to the risk diversification of the asset pool. If the pool is undiversified, then the subordinated tranche enjoys an artificially low cost of leverage and the senior tranche is undercompensated for risk. As expected default correlation goes to 0% (the pool consists of assets whose risk of default is not correlated with the risk of any other asset), leverage becomes plentiful and cheap.

Conversely, if expected correlation goes to 100%, the most senior and the most subordinated tranche should offer the same credit spread, since there is no case in which the subordinated tranche defaults and the senior tranche does not default. In this case the provision of leverage becomes impossible. Senior lenders cannot exist if the market believes risk diversification is impossible.

The Subprime Debacle, the Negative Feedback Loop and the Circuit-Breaker

Banks should not have made most subprime loans. Having made them, they should certainly not have securitized them and made them into CDOs. They applied CDO technology to an asset pool, subprime morgages, whose risk was not diversifiable.

Subprime mortgages were aggregated into mortgage pools, which were tranched. Thin subordinated tranches of the mortgage pools were then re-aggregated and re-tranched, creating CDOs whose underlying assets were both extremely risky and highly correlated. The risk derived from the structure of the CDO assets: they were themselves very thin tranches of mortgage pools, with high probabilities of 100% loss. The correlation derived from the statistical uniformity among the mortgages themselves – their performance was dominated by a single variable: home price appreciation rates.

Senior lenders thus bought the top tranches of CDO portfolios with essentially no diversification, and high risk of loss. The rating agencies routinely modeled thin subprime mortgage tranches as having default correlations in the low 20% range; in the event, true correlations have proved to be closer to 100%.

When the true nature of correlation risk in subprime portfolios became apparent, senior lenders were damaged, and sometimes destroyed. The worst hurt were the most senior, who expected the debt they bought to offer low yields but to always trade at par. Among those lenders were the institutions perceived to be the most rock-solid by the markets, the money center banks, funded at the top of their capital structures with demand deposits.

The resulting negative feedback loop threatened to arrest global capital flows.

Spreads widened; spread widening leaked disproportionately into the lowest-risk asset classes; that in turn raised refinancing costs for all borrowers. Refinancing requirements became potential default risk, hitting high-quality borrowers with short-maturity liabilities, who had become accustomed to fund themselves cheaply in the commercial paper markets.

The most senior lenders globally are those who expect to have ready access to their money at par – bank depositors and money market investors. If they cease to exist, then so do the capital markets. If they are willing to lend their money cheaply, then subordinated lenders fill in under them to create the modern capital markets.

The government managed an effective circuit-breaker last summer and fall when it guaranteed the borrowings that were, objectively speaking, the safest – highly rated commercial paper and bank deposits. That action arrested, and partially reversed, the correlation contagion and catastrophic widening of low-risk credit spreads. This in turn makes it possible for more credit-worthy borrowers to refinance, reducing default risk.

In the wake of all this, we still have banks with impaired asset valuations, skittish depositors, and paralyzed lending.

The Bank as Impaired CDO

Imagine the bank as a CDO whose assets have become impaired. The bank’s equity and hybrid capital layers sit below various intermediate levels of debt, with the depositors riding on top as super-senior lenders. If the bank were to try to sell impaired assets, the clearing sale price would be insufficient to repay all the CDO liabilities, and might wipe out tranches all the way to very senior levels. Moreover, the asset buyers would have to be entities that might look a lot like other banks, and hence the price discovery mechanism might rapidly turn into a vicious downward spiral.

Where should public money come in to all this? Coming in at any subordinated level means the government is extracting leverage from the bank’s more senior debt holders and depositors – an absurd arrangement that has private investors lending money to the government to finance risky assets. Moreover, private investors lending to the government would be lending to a player who can change the rules of the game on the liability side, and can also take unilateral action to impair or inflate the asset side of the balance sheet. This has already been demonstrated in the lurching methods applied to Bear, Lehman, and Merrill; it is also looming over the mortgage market in the form of threatened ad-hoc contract abrogations, bankruptcy cramdowns, so-called foreclosure mitigation, and so forth. In this environment, capital will stay on the sidelines and spreads will remain wide and volatile.

The proper role for public money is to guarantee the most senior part of the bank balance sheet, through an expansion of deposit insurance.


How About Toxic Assets?

The theory is that while banks hold impaired assets they are paralyzed from making new loans. The idea of setting up a “bad bank” to purchase impaired assets founders on the impossibility of price discovery. The only purchaser who can afford to ignore or supplant price discovery is the government.

On the other hand, it is possible that a sale of the bank’s assets will immediately wipe out so many layers of lenders that the bank will not be around to make any new loans.

So what is to be done? The government should start with the working assumption that every large money-center bank is already a bad bank, a virtual CDO whose assets are significantly impaired. The CDO may have unexpectedly good asset recoveries over time. In the meanwhile, it probably needs to focus on deleveraging via debt repurchase and renegotiation, dividend suspensions, etc. Rather than focusing on getting the impaired CDO to make new loans, to add to its assets, use the subsidy of super-senior guarantees to encourage the formation of new lending entities. Some of them may even be banks.

Should the government guarantee toxic assets sitting on bank balance sheets? No. In any CDO, the senior tranches are already guaranteed by the subordinated lenders. The government’s guaranteeing the assets means bailing out the subordinated lenders, with the biggest bailout going to the most subordinated tranches.

To take a numerical example, assume a CDO has $70 of super-senior debt, $20 of mezzanine debt and $10 of equity. Assume its assets are worth $60 today. Guaranteeing the assets at a value of $100 is a 100% gift to the mezzanine and the equity, and worth $10 to the super-senior. It is much simpler to guarantee the super-senior directly. That means the assets are effectively guaranteed at $70, with the subordinated lenders bearing the first $30 of losses, as they are supposed to do. The government can spend the other $30 of guarantee money on super-senior liabilities of new CDOs with better-quality assets.

Bailing out a subordinated lender is the same as buying the lender out at par, putting the government in a subordinated position, where the government doesn’t belong and from where it does active damage to the capital markets. The government should act at most as the ultimate super-senior lender – extending deposit insurance guarantees for new banks making new loans.