Thursday, June 28, 2007

Marks in the Sand

Felix Salmon put up a nice commentary today on an article in today's Financial Times which delved into the mess surrounding collateralized debt obligations (CDOs) backed by sub-prime mortgages. In it, he identifies four underlying risks in the current CDO situation and then correctly concludes that we cannot predict a priori that they will combine to push the market off the rails:
Subprime defaults can, in theory, pass through into defaults on CDO tranches. That, in turn, can, in theory, trigger CDO liquidations. That, in turn, could mean the amount of liquidity in the credit markets drying up. And that, in turn, will mean that subprime borrowers find it much harder to refinance – thereby increasing the chance that they will default.

But while all the risks are real, the linkages between them all are far from clear, and the different risks don't necessarily cascade onto and exacerbate each other in this way. They might – or they might not. If investors turn out to have reasonably strong stomachs, they might not want to liquidate at prices well below their entry points. And CDOs themselves, even the ones based on subprime mortgages, might not default nearly as much as homeowners. And without the passthrough mechanism of risks two and three, the vicious cycle loses a lot of its teeth.

So there is cause for concern, to be sure. But there isn't cause for panic.

This is correct as far as it goes. But it misses an important aspect of the puzzle which the FT article authors point to in their article. This is the fact that, to date, the CDO "market" has been largely theoretical:

[U]nlike stocks listed on an exchange or US Treasury bonds, CDOs are rarely traded. Indeed, a distinct irony of the 21st-century financial world is that, while many bankers hail them as the epitome of modern capitalism, many of these new-fangled instruments have never been priced through market trading.

Instead, products such as CDOs, which are designed to be held until they mature, have often been valued in investor portfolios or on the books of investment banks according to complex mathematical models and other non-market techniques. In addition, fund managers and bankers often have broad discretion as to what kind of model they use – and thus what value is attached to their assets.

This is the origin of the market argot "mark to model," which stands in contrast to the more traditional "mark to market."

The models used are sophisticated, but they are just that: models. Therefore, many investors in CDOs look for confirmatory quotes from third-party data sources, the banks that sold them the paper in the first place (hmm ...), and levels implied by current credit agency ratings. For reasons I will leave to you Dear Readers to discover, all of these supporting "market" quotes are flawed, and demonstrate a structural inertia that may not reflect true underlying fundamentals.

That means that on the rare occasions that instruments are traded, a large gap can suddenly emerge between the market price and its book value. This week Queen’s Walk Fund, a London hedge fund, admitted it had been forced to write down the value of its US subprime securities by almost 50 per cent in just a few months. That was because when it was forced to sell them, the price achieved was far lower than the value created with the models the fund had previously used – which had been supplemented with brokers’ quotes.

Furthermore, there is little incentive or outside pressure for a CDO investor to push for accurate security values, especially when these may be below the current "marks" in its portfolio.

But unless circumstances arise that force a market trade, valuations often remain at the investment managers’ discretion. While managers say they strive to assign honest values, these are often difficult for an outside accountant to verify, since the techniques used are invariably highly complex.

Moreover, incentives do not always encourage fair valuations: hedge fund managers, for example, are typically paid a percentage of the profits they book, giving them a vested interest in reporting a high asset valuation. At best, this means that the valuations of CDOs, for example, may often lag behind any swings in broader asset classes; at worst, this ambiguity may enable hedge fund managers or investment bankers to keep posting profits – even when markets fall.

So why is this so important, and why does it indicate a potential flaw in Felix's reasoning?

Well, the secret sauce in the CDO market stew—as in many, many other asset classes worldwide—is leverage. Many of the investors in CDOs and other structured financial products are hedge funds, and many of these have justified their hefty fee burdens to investors by using leverage to juice up returns. Buy assets yielding 6 to 8% and lever them up 3- or 5- or 9-to-1 and presto, you're delivering hefty double-digit returns to your investors, even after 2-and-20 fees. But, as anyone who has bought a stock on margin or taken out a mortgage on a home (at least recently) can tell you, leverage cuts both ways.

Hedge funds get their leverage from Wall Street, and they are not stuck with that old 50% Reg T initial margin requirement that you, me, and Aunt Madge have to pay. No, their initial and maintenance margin is determined by their prime brokers using sophisticated "value at risk" (VAR) models that usually require substantially smaller margin balances. Typically, these VAR models set comfort bands based upon some estimate of the future volatility of the underlying security (or portfolio of securities) which by necessity is heavily influenced by historical volatility. But if, for example, CDO marks in broker dealer and hedge fund portfolios have been artificially elevated—and their associated volatility has been artificially dampened—by marking to model in the absence of true market price data, what happens when someone liquidates a position at real market prices which are substantially below the mark? I'll tell you what: all hell breaks loose.

First, the prime brokers reset their hedge fund clients' margin requirements to match the current lower market prices. Bang! First margin call. Second (or simultaneously), the bank resets the input expected volatility in its VAR model to reflect the new, more volatile behavior of the securities. Bang! Second margin call. All of a sudden, a hedge fund that was sitting fat and happy with a portfolio of nicely behaved CDOs on its books is looking at a huge margin call and usually no way to meet it without liquidating securities. Oops. Fire sale, look out below.

Furthermore, if the broker dealers are holding meaningful inventories of CDOs and other structured products on their books as well—which is often the case—a general decline in price levels will put their own highly levered balance sheets under pressure, which has the knock-on effect of dampening their appetite for buying such securities from their clients and otherwise providing market-making liquidity. And, other things being equal, less liquidity means lower prices. And lower prices ... well, we saw what those caused in the preceding paragraph.

Maybe this is why Wall Street pulled back from the brink when it came to liquidating collateral in the Bear Stearns' CDO funds:

So when Wall Street creditors last week threatened fire sales of CDOs seized from the stricken Bear Stearns funds, thus creating a market price for them for the first time, they also threatened to create a wider shock for the system. Fire sales rarely realise anything close to the previously expected value of assets. But if these deals went ahead, they would provide a legitimate trading level that would challenge current portfolio valuations.

In other words, they saw just how deep the abyss was, and they chose to break for tea instead.

A similar risk dynamic exists in any market which consists of relatively illiquid, uncertainly priced securities and which has a material portion of its participants indulging in margin leverage. Now, I do not claim to be a market expert on CDOs, but I think I can assert without fear of contradiction that a material portion of investors in CDOs are in fact hedge funds which employ leverage as part of their investment strategy. And all it takes is a few such forced sellers to turn a market decline into a comprehensive rout.

This is the trick. Even if a hedge fund manager is convinced that the market is underpricing CDOs right now, he will have to sell in order to meet his broker's margin call no matter how strong his stomach is. That is one of the instructive lessons from the Long-Term Capital Management implosion: Meriwether and his colleagues were convinced all throughout LTCM's long collapse that the market was irrationally selling into all their positions. They did not want to liquidate because they knew they were right about their portfolio's underlying value. Guess what? In retrospect, they probably were right. But that didn't help them avoid the forced liquidation and takeover of LTCM to meet their Wall Street financiers' margin calls.

A strong stomach won't help you when you've borrowed a lot of money and your broker wants it back.

© 2007 The Epicurean Dealmaker. All rights reserved.