Resident Fellow Alex J. Pollock

The subprime mortgage lending bust, now in process, displays all the classic results of recurring credit overexpansions. Such credit celebrations are based on optimism and a euphoric belief in the ever-rising price of some asset class, in this case, houses and condominiums. They are inevitably followed by a hangover of defaults, failures, dispossession of widows and orphans (already the subject of congressional interest), and the risk of a late-cycle regulatory reaction which drives the market further down.

In the general pattern, nothing changes. As the great student of economics and finance, Walter Bagehot, observed in 1873: "The mercantile community will have been unusually fortunate if during the period of rising prices it has not made great mistakes. Such a period naturally excites the sanguine and the ardent. Every great crisis reveals the excessive speculations of many houses which no one before suspected."

In the period of rapidly rising house prices, the mortgage lending community, some of its members no doubt sanguine, ardent, and enjoying big commissions and bonuses, made some significant mistakes.

Now we have the failure of numerous subprime mortgage banks, delinquencies and defaults accelerating to unexpectedly high levels with unexpected speed, and growing political-regulatory risk for the industry. World-class HSBC Bank, with its recent announcement of subprime credit problems and related management changes, is a nice example of a "house no one before suspected."

How could all this happen with the massive and wonderful computer power manipulating all the data with the complex models built by mathematical experts we now have? The former CEO of Household International, which was bought by HSBC, is said to have bragged that his operation had 150 Ph.D.s to model credit risk. Might this make us wonder about the huge subprime securitization market, tranched and sold to yield-hungry investors based on the models of the credit rating agencies?

Let us consider Moore's Law of Finance (named for my friend Mike Moore of Hillenbrand Partners): "The model works until it doesn't." Perversely, the more everyone believes the model, and the more everyone uses the same model, the more likely it is to induce changes in the market that make it cease to work.

It appears that in this cycle the market became enamored with the statistical treatments of risk, while the most important issue is always the human sources of risk. These human sources include short memories and the inclination to convince ourselves that we are experiencing "innovation" and "creativity," when all that is happening is a lowering of credit standards by new names.

For example, with the spread of "stated income" loans, the disastrous previous experiences with "no doc" and "low doc" loans seem to have been forgotten. Such loans are a notable temptation, or even invitation, to a little exaggeration, let us call it, in order to facilitate the dream of buying the house whose price will always keep rising.

Human elements of risk also include optimism, gullibility, short-term focus, genuine belief in momentum or the extrapolation of so-far successful speculation, group psychology or the lemming effect, and inevitably in some cases, fraud.

We should not be surprised that as optimism increases, so does the incidence of fraud. As Bagehot further observed: "The good times of too high price almost always engender much fraud. All people are most credulous when they are most happy; there is a happy opportunity for ingenious mendacity. Almost everything will be believed for a little while."

How unhappily true. In the modern classic Manias, Panics and Crashes, Charles Kindleberger and Robert Z. Aliber expand upon this theme: "The propensity to swindle grows parallel with the propensity to speculate during a boom. The implosion of an asset price bubble always leads to the discovery of fraud and swindles."

If history is any guide, we can expect the same from the implosion of the subprime mortgage boom. On the same basis, we can also expect the usual political overreaction, which may result in the Sarbanes-Oxley Act of mortgage lending, or some similar mistake.

The subprime boom cannot be discussed without considering its driving financial engine: securitization of subprime pools through tranched, senior-subordinated structures based on models. The lower tranches of subprime MBS are highly leveraged to credit risk. But they are often gathered into CDOs and further tranched, thus creating securities hyperleveraged to credit risk.

Where did these exceptionally risky tranches go--who is holding them? What will happen to them and how will they react as the subprime mortgage bust proceeds? These essential questions must put us in mind of Stanton's Law (named for my friend Tom Stanton of Johns Hopkins University): "Risk always migrates to the hands least competent to manage it." This is because the more competent can manage their risk by passing it on to the less competent. But there is another possibility: "when genius fails," the extremely clever may believe too much in their own models and cleverness, then find out they had much more risk than they thought.

It seems we have some interesting times ahead.

Alex J. Pollock is a resident fellow at AEI.