Cash Collateralized debt obligation (CDO)
by David Harper, CFA, FRM, CIPM
Learning Outcome (LO 54.2)
- Explain the structure of a typical cash collateralized debt obligation (CDO), including the use of a special purpose vehicle.
A CDO is an asset sale and a risk transfer into securities
A cash CDO is an asset sale. The originator (e.g., a bank) sells credit-sensitive assets to the special purpose vehicle (SPV). There are several motives for the bank to sell its assets. These motives include, but are not limited to: to remove the assets from the bank's balance sheet, to receive cash ("to monetize" the assets), and to transfer risk. The risk transfer is an essential feature. Consider the following CDO:
In this case, we illustrate mortgages as the underlying collateral. But the collateral (the credit assets) can be commercial mortgages, corporate bonds, preferred stock, among several other securities types including notes issued by other CDOs! Note the fundamental mechanics above:
- The bank (not shown) sells a portfolio of credit-sensitive assets to the special purpose vehicle (SPV) in a so-called true sale.
- How does the SPV pay for the assets? The SPV issues securities to investors (facilitated by underwriters, who broker the securities)
- The blue line above illustrates the risk transfer. Consider the investors: they purchase securities and, in exchange, they receive the principal and interest (P&I) on the underlying credit-sensitive assets. In general, they now collectively bear the credit risk.
- The green line above illustrates the asset monetization. The bank has sold the assets for cash. So, the bank achieves three things: removes assets from the balance sheet, receives funds, and transfers credit risk to the investors.
- The holder of the junior tranche has a risky and residual claim: this investor is selling default insurance (as the originator is essentially re-insuring the assets) but gets paid the excess spread
Risk transfer
As the sub-prime debacle proves, the risk transfer is not simple. Investors purchase securities that represent different tranches (French for 'slices,' says John Maudlin). The credit assets have been repackaged into tranches, essentially. The tranches are arranged in a sort of totem pole hierarchy. Defaults "flood the bottom," so to speak, at first and then rise up the totem pole. At the bottom is the tranche variously called the junior tranche, the equity tranche, the first loss piece and famously the toxic waste tranche. This junior tranche absorbs the first defaults and therefore enjoys a higher yield.
The senior tranche is often rated AA or AAA. This is partly a function of the underlying basket of assets, but it is more a function of correlation (read: diversification). The less correlated the assets, the less likely defaults will reach the senior tranche. Of course, a student of risk will recognize at least two challenges right-away (get more detail in this excellent primer from the knowledgeable Accrued Interest):
- Correlations are dynamically time-varying. The spike at the worst times, when it really matters, and render the diversification plank irrelevant
- To the extent the assets are illiquid, there is already significant model risk. The parsing (into tranches) of illiquid assets creates model risk on an epic scale.
Special purpose vehicle
The learning outcome asks about the "special purpose of the SPV." The purpose of the SPV is to enable a true sale of the assets away from the originator (bank) to a third-party. In principle, the idea is very simple: if the bank still controls or owns the assets, it should consolidate them on its balance sheet. As assets on the balance sheet, they decrease return on assets (ROA) (all other things being equal) and the bank must hold capital against them under the Basel Accord. But the details are not so simple.
FASB Statement No. 140 (FAS 140) lists four criteria for a true sale:
- The SPV is bankruptcy remote and the assets in the SPV are sufficiently isolated from the originator to survive Chapter 7 or Chapter 11 bankruptcy.
- Permissible activities of the SPV must be significantly limited, must be specified at deal inception/incorporation of the vehicle, and can be changed only with approval of a majority of interest holders other than the originator or its affiliates or agents.
- The originator must surrender “effective control” over the assets
- The SPV must have the right to pledge/resell/exchange the assets acquired from the originator. The buyer must have a “perfected interest” in the acquired assets.
So FAS 140 was violated by Enron, for example, because they did not consolidate securitized assets that were effectively controlled by Enron. This famous violation, in turn, gave rise to FIN 46R (the "anti-Enron" rule) which tried to close the loophole.
In summary
One of customers recently asked, "Isn't the purpose of securitizing to get assets off balance sheet?" To which, the answer is yes. But it is not the only motive. We have mentioned three motives here, in the context of a CDO:
- Transfer risk: the bank wants to transfer credit risk (default risk) to the investors
- Monetize: the banks wants to receive cash
- Shrink the balance sheet: the bank wants to shrink its balance sheet which has several benefits. Some of those benefits may be real and economic, but some are "merely" motivated by accounting (how can we avoid consolidation?) rules and regulations (how can we minimize capital required by Basel I/II?)
The investors have motives, too, of course. They are earning higher yields. The junior tranche holders take high yield, high risk positions. In the case of sub-prime mortgages (i.e., CDOs collateralized with subprime mortgages), we might say junior holders are selling a sort of catastrophe insurance. The problem in the case of subprime, is that the defaults are progressing up the totem pole into higher-quality tranches that are supposed to be buffered by diversification of the basket.
Comments
Hi David,
On the wave of the losses suffered by Merill Lynch, where exactly would one place Merill Lynch on the diagram above (Cash Collateralized Debt Obligation), since it’s this same Cashflow CDO that was the bane of the CEO, Stan O’neal?
Thanks for your usual insightful explanations.
Telon
Hi Telon,
I don’t know, really. On the earnings call they deflected queries about details of their CDO exposure. The above diagram is meant to introduce CDOs, so (1) it simplifies; e.g., the originator on the left often retains the junior/equity tranche on the right, but i did not want to clutter an introduction, and (2) the cash CDO is just the beginning of the variations
My guess is Merrill is *everywhere* on the diagram, but the real pain was in the senior tranches (upper right corner). Clearly they are: an originator, underwriter, certainly an investor (holder of tranched securities) and, in regard to their investor status: they explained on the call, they hold senior, mezzanine and junior tranches. It appears the “debacle” write downs are not where you’d expect them—in the junior tranche. Sure, they are worthless, but I doubt Merrill had unreasonable exposures in these ‘risky by design’ toxic tranches but, instead, the highly-rated tranches are collapsing as they sit on a house of cards.
The one party that doesn’t suffer too much in the CASH CDO above is the originator, they are fully funded. So, they probably are not there too much…
My guess is Merrill is NOT big in the cash CDOs (as above). Cash CDOs start the FRM introduction because it emphasizes risk, and the cash CDO is motivated by the originator. The originator is looking to get loans off the balance sheet. Merrill likely has a much larger business in investor-motivated CDOs, esp synthetic CDOs and so-called bespoke CDOs. So, instead of starting with a balance sheet (left, above), they are going to investors and saying/selling “we will synthesize a CDO for you.” They synthetize with credit default swaps (CDS) and more specifically with synthetic arbitrage CDOs. Where cash CDOs are a risk transfer vehicle; these synthetic vehicles start with chasing yield. So, surely they crafted a lot of bespoke CDOs and, as underwriters, retained tranches themselves on the same. It would be nice to know, in my opinion, because the motivation for the CDOs (i.e., asset owners originating or speculation-seeking-investors) would help explain whether and how reckless was the strategy.
I hope that helps, I’d love any references/links to people closer to the particulars…
Thanks very much for the response.
“collateralised debt obligation default” is repeatedly mentioned. What causes collateralised debt obligation default, that is the criteria?
William,
It matters whether it’s a cash or synthetic CDO and, if synthetic, to the extent it’s funded. The above illustrates a basic cash CDO (the assets are sold the SPE). “Defaults” manifest in the cash flow waterfall - the logic/rules that govern the splicing the cash (P&I) from the securitized assets - the investors are participating in the cash flows indirectly. It’s the waterfall logic that determines their cash received; the junior tranche holder just doesn’t get paid when the cash flow isn’t received.
In a synthetic CDO (the SPE is writing credit protection), the waterfall starts with any CDS obligations: the SPE first may payout on any defaults before cash, according to the waterfall, goes to the investors. This is unfunded, and so the contractual definition of ‘default’ is critical (it rules the top of the waterfall). I don’t have numbers, but it is my understanding most follow ISDA (http://www.isda.org) which is standardized but also flexible so particular default criteria with ISDA may vary (e.g., grace period).
Leave a Comment