Summer Rerun: CDOs: The Ticking Time Bomb

This post first appeared on November 10, 2007

The equity markets seem to have finally realized that conditions are ugly in the credit markets, due to get uglier, and the mess will pull down the real economy.

And the bad news continues.

The dollar index fell to a new low. Wachovia said the value of its subprime securities, largely “super senior” tranches of CDOs, fell $1.1 billion and it was witnessing sharp falls in housing prices in some areas of the US.

We had questioned the optimistic assumption of a few weeks ago, that investment banks were taking deep enough writedowns to put their woes behind them. How can you possibly do that unless a market is at or near bottom? Wall Street analysts have indicated that they expect further writeoffs at the Wall Street firms. Consistent with that view, Bank of America and JP Morgan said in their SEC filings that the fourth quarter would bring more writedowns. In particular, Bank of America noted that it expected “significant dislocations” in the market for CDOs. Fannie Mae reported a loss double that of a year ago. Capital One said it was seeing a faster-than-expected increase in credit card delinquencies.

But what I believe will be the largest single source of trouble in this sorry credit picture is coming to the fore: collateralized debt obligations.

CDOs are structured investments that contain a variety of assets, often tranches of residential mortgage securitizations (often the BBB or BBB- portion), commercial mortgages, pieces of collateralized loan obligations (LBO debt), and sometimes whole residential mortgages, There are also CDO squared (CDOs made from CDOs) and CDO cubed, and synthetic CDOs (CDOs constructed from the cashflows of credit default swaps written as credit enhancement for other CDOs). Because the assets in CDOs are heterogeneous, and the mix is particular to each deal, the structures vary tremendously as well.

In other words, CDOs are arcane and hairy paper. Ever since subprimes went wobbly, CDOs looked primed for trouble. For more background, see here and here.

There is no method for reporting on CDO issuance or on subprime loans outstanding (there are classification issues for subprimes, and that is compounded by the fact that many of the originators were mortgage brokers who had no reporting obligations). But to give a sense of relative importance, the estimates we’ve seen of the size of the subprime market range from $1 to $1.3 trillion (rated subprimes are $565 billion). By contrast, the estimates of CDOs outstanding are all over the map, but the ones we find more credible range from $2.0 to $3.9 trillion (we have also seen smaller estimates, but when the Financial Times, which has done far and away the best job of covering this market, reports that global CDO issuance in 2006 alone was $2.6 trillion, we tend to put our faith in the bigger numbers.

The reason for the disparity may have to do with what was counted. There are both passive CDOs, which are more like a conventional asset backed securities deal, versus managed CDOs, in which funds are raised on a blind basis and given to a CDO manager, who then buys the CDO holdings and can and does trade the holdings to improve results It is also an open question as to whether the variation in estimates is due to the inclusion or exclusion of synthetic CDOs, whose assets are credit default swaps written on CDOs. One source reported that synthetic CDOs were nearly 1/3 of total CDO issuance in the first quarter of 2007.

As readers probably know all too well, CDOs had been valued using models that the rating agencies had developed solely for credit rating purposes. Investors had carried CDOs on their books at more that they were probably worth, at first out of ignorance, because they didn’t realize the subprime paper they held was weaker than it had been historically, and later out of convenience (why recognize losses if you don’t have to and no one else seems to be doing so either?). And the investment banks were similarly able, in the absence of price discovery, to carry the CDOs on their books (they wound up owning them by not being able to sell out the full amount they created) without marking them down too much.

Those days are rapidly coming to an end. CDOs were the big culprits in the Merrill and Citigroup writedowns. And we have two developments coming to a head on the same time frame. The first is that new accounting rules gives companies far less latitude in how they value this paper. As the Financial Times explained it:

They are the “buckets” into which financial statement preparers must classify financial assets under FAS 157, a new US accounting standard for financial years beginning in November…

At the top of the bucket hierarchy is Level One, involving assets with prices quoted in active markets, such as mainstream stocks. Level Two contains less-traded securities and uses prices for assets very like the one being valued.

At the bottom lurks Level Three, assets with “un observable inputs”, meaning their value is calculated via a series of assumptions. Most collateralised debt obligations end up here.

While these categories may be familiar to many readers, what is not as widely know is that another rule, FASB 159, pushes institutions to put positions into the lowest bucket possible. Thus, no phony-baloney Level 3 valuation if there is a way to come up with a gridded or extrapolated Level 2 value.

The second development is that markeplace changes are forcing the revaluation of CDOs. Having first gone through re-rating subprime bonds, they are now tackling CDOs, and downgrades will force commercial banks, investment banks, pension funds, and other holders to recognize losses.

More worrisome is the rerating process is putting CDOs into default. As the Financial Times explains:

Fire sales of mortgage assets from complex debt vehicles began in earnest after the trustee of a $1.5bn complex debt deal managed by State Street Global Advisors started liquidating its portfolio.

Ratings downgrades for mortgage securities have pushed a clutch of such deals into default. Trustees have issued default notices for more than 14 collateralised debt obligation deals in recent weeks, representing securities with a face value of more than $10bn.

A default means the most senior investors in the CDO can liquidate the underlying assets to get their money back. Analysts say more deals are on the brink of default.

$10 billion is a pretty small amount by bond market standards, but the process of liquidating CDO assets is going to weaken prices even further in the mortgage securities market, and financial institutions will have to remark their positions in line with these new, lower values. And there is every reason to believe that CDO liquidations will accelerate.

Another problem that investors are discovering to their dismay is that CDOs have a tremendous amount of embedded leverage. That means, in lay terms, that comparatively small changes in the cash flows from their underlying assets have a large impact on their value. That is why, when the rating agencies re-rate the CDOs, the downgrades can be staggeringly large, not a mere one or two rating grades, but 11, 13, even an unheard of 18 grades.

And this casts the ratings into even further doubt. What does a triple A mean if, as one wag put it, you can go to lunch and due to a downgrade, you can come back and find that it is now junk? That simply doesn’t occur with other instruments in the absence of fraud. So the problem isn’t simply that the rating agencies might have come up with overly high ratings when the CDOs were issued because they had overly optimistic estimated for likely defaults. You have the further problem that any downgrade is likely to be dramatic. So even if you believed that your CDOs AAA was really an AAA, you’d still value it less than AAAs on other securities because the downgrade slope is so steep.

Now a fair question is whether the dim view of CDOs is overdone. According to Bloomberg, Morgan Stanley argued that case in a recent research note:

More than $350 billion of collateralized debt obligations comprising asset-backed securities may become “distressed” because of credit rating downgrades, Morgan Stanley said in a report today.

“The pace of ABS CDO downgrades will pick up significantly over the next few weeks,” wrote analysts led by Vishwanath Tirupattur in New York. “Given the degree of market dislocations and the potential size of the market, there is clearly an opportunity for attentive investors.”…

Bonds are considered distressed when investors demand yields at least 10 percentage points above similar-maturity Treasuries. Morgan Stanley said many of the CDOs of asset-backed securities sold in recent years are trading on an interest-only basis, signaling investors expect to receive little of their original principal back.

I for one would not try to catch that falling safe. While the higher-rated tranches of many CDOs almost certainly have some value, the problem is that these instruments were sold way beyond their natural market due to misplaced confidence in their ratings. Consider how small the market would be if the paper carried no rating.

Moreover, most fixed income investors are professional money managers subject to performance pressure. Even if one of them thought certain CDOs were cheap, he’d still hesitate to buy them out of concern that the prices would drop further before they rallied, and he’d show losses on those positions (and worse, have to explain to his boss and/or clients why he bought CDOs). Very few are going to step in until the market appears to be improving, and between now and then, we are likely to discover there is way too much CDO paper relative to who wants to own it now.

And let’s consider an ugly factoid. In its third quarter results, Merrill wrote down its CDO holdings by an estimated 30% or more. These were reported to be almost entirely AAA rated. This does not allow for either the further deterioration, nor the fact that downgrades are proceeding, which will impair value even more.

Take an estimated $3.0 trillion in CDO outstandings. Apply a mere 25% loss to them. That’s $750 billion, roughly four times greater than mainstream estimates of subprime losses.

Now admittedly there is subprime paper included in those CDOs, so there is some double counting, but the CDOs’ subprime holdings reportedly contain mainly the BBB/BBB- tranches, which is a portion of the value of the initial RMBS securitization.

Houston, we have a problem.

Update 11/13, 3:30 AM: Thanks to the Financial Times’ MergerMarket blog (hat tip Felix Salmon) we have some better data on CDO market prices, and how much Merrill wrote down various grades of CDOs:

However, AAA rated subprime CDOs currently trade from the high single digits on junior tranches to 60% of face on super senior tranches, according to a sellsider and a buysider…..

Merrill Lynch in the third quarter discounted its own super senior ABS CDO holdings by an average 19%, while mezzanine AAA notes were written down by 37%.


Print Friendly, PDF & Email


  1. marc fleury

    In retrospect by november the bomb had blown up. I don’t mean to downplay the issue, but CDO valuation can be considered l3 assets (long dated, illiquid, poor market pricing) and the 60% haircut in the market for mezz tranches was a mark to market issue for many banks. This led to lack of overnight funding and the well known roubini dynamics that brought the investment banks as they knew it, to its knees. However the trigger for the market was also a tight illiquidity condition. That in turn was triggered by naked CDS which imploded in aug 2007 or 4 mo before the article. It was that cash drain that created illiquid conditions in the market and PRECIPITATED the downfall.

    I think we would all do well to remember that the dynamics of the fall are just important as the fall itself. I am not arguing that the CDO and the market shouldn’t fall, they should but how fast is the question. Markets are irrational but when they trigger wide balance sheet bankruptcy, one has to pause, as we did, and think about FAS 157 and mark to market.

    In retrospect the liquidity measures introduced by the FED, to backstop the slide in residential CDO markets for example, was right on. It does not prevent the fall, eventually the market will relax to historical levels but not too fast is the key… maintaining liquidity is in fact a FED mission :)

  2. Siggy

    Interesting point of view.

    Can anyone explain to me why anyone would buy a CDO? A CDO is comprised of income streams attributable to first lost tranches of other securities. In that their expected value is close to zero, what price is worthy of such paper? Ignore the ratings, the agencies had no experience in rating these securities. What ledgerdemain permits the awarding of AAA for securities that have such a tenuous claim on questiuonable income streams?

    You’ve got to be really stupid to buy into the CDO mania. The collapase was set in stone the day the damm things were sold or held on the books of the issuing institution.

    The proper resolution of this mess should have been the bankruptcy court. Would that have created global financial chaos? Quite probably so, and that chaos may well have been and come to be what is necessary to resolve the insolvency of so many over-large and criminally managed institutions.

  3. Jackrabbit

    And check out this doozy:

    From Bloomberg: “Fitch cut its credit ratings on Cayman Islands-based Axon Financial’s commercial paper and medium-term notes to “D” in November 2007, about eight months after granting top grades [AAA] to the debt as [i.e. when] the SIV was created.” (Note: other rating agencies also downgraded multiple notches about the same time.)

    57 percent of Axon’s SIV portfolio was US Residential MBS, and 6 percent was ABS CDOs (CDOs of ABS – mostly MBS). So almost 65% of this AAA-rated SIV was MBS (The high rating was because many were AAA-rated tranches). As reported by Reuters, “Moody’s said Axon’s capital net asset value had fallen to 39 percent on Oct. 18 2007 from 96 percent on July 27 2007.”

    A new record? From 100 cents to 39 cents in 7 months; from ‘AAA’ to ‘D’ in 8 months.

    TPG-Axon created the SIV (Axon Finanical Funding, managed by Axon Financial Services) with roughly $275m of equity – 5 months later they wrote down this equity to ZERO. Despite this loss, Axon claimed to have made money trading subprime during this time by betting against subprime (sound familiar?). See here or Google it (available via Google cache, if you don’t subscribe).

    TPG-Axon Capital Management, is a hedge fund led by Dinakar Singh, who previously ran Goldman Sachs’ in-house trading unit.

    1. Jackrabbit

      Soo… given what we know about other strategies (Paulson, Magnetar) for profiting from a subprime collapse, the question is: did TPG Axon have such a strategy? For example, they could’ve bought the equity and shorted the capital notes. As a trading strategy it is brilliant but it 1) added fuel to the fire, and 2) begs the question of disclosure to (sophisticated) investors (I believe that Legg Mason and the State of Florida lost hundreds of millions each).

    2. Jackrabbit

      I think the question of whether TPG-AXon had found a way to play the impending subprime implosion is important for public policy. The greater the number of investors that anticipated the implosion, the more difficult it is to say that no one could foresee it. For the regulators to let themselves off the hook by saying that no one could’ve foreseen the bubble and subsequent bursting is unacceptable.

  4. charles

    A question from Europe: Was the ‘level 3’ FASB implementation
    due to the Treasury, as later on with the current Treasury
    ‘s ‘mark to market-mark to no-market value’ rule ? In other words, was the rule introduced as the tree to ‘shade’ the
    forest-or rather its imminent ‘deforestation’? Nessim Taleb in
    the update of his opus writes: ‘Ín the 2000s the banks took over the government.’
    European accounting is a little different, safe for Spain
    where the Banco of Espana changed the rules for NPLs to ease ‘zombie’ banks ?

  5. Noel

    To the Author can I have permission to use some of the information from this above post as long as I post a link back to your blog?

Comments are closed.