Blast from the past

From 2007, predicting a number of the problems of 2008 and beyond:

Phase One

Even within the last year, you may recall driving down the highway listening to one of the many mortgage ads on the radio and hearing the phrase: “No income verification.” Perhaps you thought: “That seems a little odd. I wonder how they do that. They must have very sophisticated computer models.”

Well, the phrase is still around, but for the most part the loans aren’t. And therein lies a tale. In the olden times, say, 20 years ago, you went to the bank to get a mortgage when you bought a house. The loan officer went over your tax returns, credit report, bank statements, proof of down payment and all the rest, and the loan committee decided if you were creditworthy. That all ended when mortgage securitization was invented and then became widespread, expanding massively in this decade, as interest rates reached historic lows.

It is much easier to use computer models than credit judgment. And computer models showed fascinating facts about mortgage holders and defaults. They showed consistent correlations between different credit ratings and mortgage default rates, for example. So using various inputs of this sort, computer models of large numbers of mortgages showed that only a small portion of the mortgage holders default and most mortgages were as good as gold. This allowed smart fellows to divide up, say, $1 billion in mortgages, into various groups, or tranches (see Allan Sloan’s article).

The tranches then got ratings from the credit agencies, well known entities with long histories like S&P and Moody’s. Based on various metrics gleaned from statistical history, large numbers of these mortgages were rated AAA, and lesser mortgage pools got lower ratings, etc. So then the mortgage tranches could be sold to different groups who wanted AAA paper at a low yield, or riskier paper at higher yield, etc.

But the computer models were wrong. They made, in essence, the same mistake that Mike Milken made in his analysis many years ago of the low default rates of junk bonds. It was true that junk bonds had historically low default rates, but of course that was not an Iron Law. In fact issuing vast numbers of junk bonds changed the very nature of the game, attracting a group of wise guys who sought to exploit the fact that investors had substituted a computer model for common sense. So it became with subprime mortgages, when a flood of new subprime borrowers changed the game and created the certainty that future default rates would vastly exceed those from loans made in a previous time of prudent lending.

It wasn’t just the issuers of these Collateralized Mortgage Obligations (CMO’s) or the rating agencies who were using these bad computer models. The buyers of these securities were as well. And just who were the buyers? The buyers included the traditional banks and insurance companies here and abroad of course, but also included sophisticated trading operations of investment banks, as well as the relatively new category of investment institution called hedge funds.

One of the ironies of our age is that a great many of these hedge funds are not in fact hedged in any meaningful sense. That became clear enough this summer. Though there were some earlier headlines, around the beginning of August, the markets seemed to focus on the subprime matter, all at once. And all at once the computer models used by many of the smartest investors around began to shout SELL simultaneously. It is hard to forecast discontinuities, which are inherently difficult to anticipate, and the computer models did not do so. As a result, many hedge funds had to sell assets at the same time to raise liquidity, which played havoc on the equity and debt markets.

Of course not everyone was caught by surprise, though many of the smartest firms were. Morgan Stanley got caught. Its CFO said, referring to the extensive computer modelling that the firm had done, “no stress model in the world would ever have had it.” Goldman Sachs, previously one of the prime purveyors of subprime paper, fared much better. It apparently was well aware of what it was shoveling into the marketplace, and was not fooled by computer models. Goldman made money from precisely the same circumstances that caused Morgan to lose money.

Phase One of the credit crunch appears to be winding down at the moment with the White House and Treasury proposing some sort of subprime bailout. But our story doesn’t end there.

Phase Two?

Many of the same structural conditions of the mortgage market have made their way into the corporate debt market as well. We mentioned CMO’s above. There are also such things as corporate Collateralized Debt Obligations or CDO’s.

The same dynamics of substituting computer modeling for credit judgment have come to apply in the corporate debt arena as well. That is one reason why so many highly leveraged deals have been able to be sold to the marketplace, until this summer. As the ex-CEO of Citigroup said in July, among many other foolish things, Citigroup was “still dancing” at the party of loan syndication, where “liquidity rushes in” to fill gaps in markets. He’s not the ex-CEO for nothing.

In the old days, banks made loans to creditworthy customers. If the loans were large, they were syndicated among banks, with one bank acting as the “lead.” The bank group was stable and often of long standing. People knew each other, often for many years. If difficulties arose in a credit, they were addressed by knowledgeable people who understood the company and its industry. With securitization, some of these dynamics change.

We described how different tranches of mortgage securities were sold to hedge funds and other institutions who wanted various yield versus risk trade-offs. The corporate debt market adds a couple of highly significant wrinkles to this concept.

Of significant importance is that a specific tranche of a debt instrument may wield a great deal of power in dictating the affairs of a company. A classic example is that in bankruptcy, according to the priority rules, the first impaired creditor class becomes the equity in a reorganized company. This has given rise to a mini-industry, the loan-to-own hedge funds, institutions that deliberately buy themselves into (at a deep discount) the debt of a troubled company, with the idea that they will ultimately control the company and make outsized returns.

And that is just one of the tricks of the trade of these activists and wise guys. Another example would be a hedge fund owning part of a tranche of debt that could approve or deny the waiver of a particular debt covenant for a company in default. The fund might make it known that it will not consent to a waiver, meanwhile clandestinely buying up major pieces of debt at a lower level in the capital structure at a huge discount, and then approving the waiver. Immediately the fund could make a profit on its manipulated investment in the lower parts of the capital structure. We haven’t even scratched the surface of the funds’ creativity with these examples.

While insider trading has made up a minor part of the equity markets for many years, as a result of enforcement activities, it would appear to be routine in some debt instruments, where trades are not public, sometimes do not involve “securities” at all, and often take place among institutions who have agreed in advance not to sue each other over material non-disclosures. Though the SEC has prosecuted some insider trading cases involving debt since 1943, these are the exception, not the rule. Thus hedge funds of a certain mindset have every incentive, and few structural limitations, to try to manipulate certain markets.

Finally, we want to mention yet another instrument that has arisen from a reasonable need, but would appear to offer a further opportunity for abuse by the wise guys. That is the credit default swap, an instrument whose spreads have widened considerably over the course of 2007. The value of a credit default swap varies of course along with the underlying credit of the company whose credit risk is being hedged. That much is obvious.

However, the value of a credit default swap also can vary as a result of the assessment of the health of the credit markets generally. That is because a credit default swap is merely a contract between two financial counterparties, and is only as good as the health of those counterparties themselves. As Warren Buffett observed:

“Unless derivatives contracts are collateralized or guaranteed, their ultimate value also depends on the creditworthiness of the counterparties to them. In the meantime, though, before a contract is settled, the counterparties record profits and losses — often huge in amount — in their current earnings statements without so much as a penny changing hands. The range of derivatives contracts is limited only by the imagination of man (or sometimes, so it seems, madmen).”

The general point to be made about the loan-to-own hedge funds, the wise guy traders in distressed debt, and credit default swap speculators is this: there are tremendous fortunes to be made by short selling the entire financial system. There are people and institutions who benefit if credit markets seize up, banks default on their obligations, or companies file for Chapter 11. One possible reason for the talk of doom and gloom, and periodic rumors of panic, in a period with objectively strong economic statistics is that there are investment funds who have shorted prosperity and have gone long doom and gloom — and they have every incentive to make that bet profitable.

Leave a Reply