Adoption of credit derivatives technology has been key to the explosive growth in collateralised debt obligations. In 1990 the CDO market consisted of roughly US$2.2bn in rated securitisations. In 1997 the US$64bn rated CDO market consisted chiefly of securitisations of cash assets. Growth seemed impressive, but the biggest surge was yet to come due to the growing credit derivatives market. In 2002 CDO global issuance was almost US$270bn, and YTD 2003 estimates of visibile issuance were at around US$370bn at the end of September, already exceeding 2002 issuance by 37%. Synthetic CDOs, or securitisations incorporating credit derivatives technology to transfer asset risks and cash flows, now make up more than 75% of CDOs.

The unfunded nature of the credit default swap and the ease of manipulation of cash flows allow more flexibility in the creation of deals. Synthetic deals can be created in greater size than cash deals, and avoid the currency and interest rate risk of most cash based CDOs. Synthetic deals pose other challenges for investors, however. Close attention must be paid to the structure of the deal and the cash flows involved.

The viability of the synthetic ‘arbitrage’ using exclusively investment grade collateral is a huge advantage over cash CDOs, which require some non-investment grade collateral, in the current credit environment. The super senior tranche, which prices well below a typical Triple A tranche and which makes up more than 80% of the synthetic CDO, is a major driver of the economics of the synthetic CDO. The feasibility of using a portfolio exclusively comprised of investment grade collateral allows increased leverage of the equity tranche, providing a second major driver of the economics of synthetic CDOs.

The accompanying table shows the order of magnitude of the difference in the typical tranching of a managed cash versus a managed synthetic CDO.
Notice the size of the super senior tranche. It is 86.5% of the synthetic deal. Super seniors have traded in a ballpark range of 6bp pa to 20bp pa in the past two years. Triple A tranches have traded in a ball park range of 45bp pa to 75bp pa in the past two years. In the case of a cash deal, the range would be expressed as a spread to three-month LIBOR. If the super senior trades at 10bp per annum, and the Triple A tranche trades at 50bp per annum, the synthetic deal has an arbitrage advantage of E1.665m per year, on the liabilities above the Triple A attachment point, even though the Triple A and super senior is a larger portion of the synthetic deal.

In the synthetic CDO, what happens to all of the excess cash freed up by the ability to issue the inexpensive super senior liability? What should happen to it?

For banks and insurance companies that invest in super senior tranches, it is important to note there is no market standard definition of super senior risk. There is also no standard means of pricing super senior risk. The super senior is the most important driving force in synthetic deals and the super senior is the largest chunk of a synthetic deal. How can it be there is no standard pricing method and no standard definition? Super seniors have been represented in many marketing presentations as ‘Quadruple A’. This is simply a misrepresentation. Rating agencies don’t recognise the existence of a super senior tranche. For instance, in the example given above, I could have created a 2% Triple A tranche and an 88.5% super senior tranche and still be ‘correct’. It is incredible, but true. The tranching is whatever the market will bear.

Given that is the case, the super senior investor may want to ask for structural protection. For instance, the excess cash flows may be diverted to a reserve account under certain conditions. At the maturity of the deal, if the cash that has accumulated in the reserve account is not needed to support the integrity of the super senior tranche, it can then be released according to the priorities stated in the deal documents.

The Triple A investor in the synthetic CDO might also consider asking for support in the form of a reserve account. The horizontal tranche underneath the super senior tranche is typically sold as if it has a Triple A rating. My personal view is that this tranche does not merit a Triple A rating. It is not a Triple A tranche in the customary sense of the Triple A tranche of a cash CDO. For instance, S&P would view this slice as having only a double A rating because, by its definition, there can be nothing above a Triple A tranche.

Suppose you were unlucky enough to have bought one of the deals that had WorldCom, and Enron, each as 1% of the deal, so that 2% of the total names in the portfolio defaulted. Chances are that other names in the portfolio were downgraded, so that the amount of subordination to maintain a Triple A rating would have to increase by more than 2%. But you know for sure that it would have to increase by at least 2%. Using this knowledge, let’s take a look at the effect on two E25m Triple A investments – one cash, one synthetic.

The two Triple A tranches are not equivalent in a non-static world. If everything else remained the same, but 2% of the portfolio defaulted, about E0.57m or about 2.27% of the E25m cash Triple A tranche would not be deemed Triple A. If we restructured to add 2% more subordination to the total deal, about E24.43m of this tranche would still be deemed Triple A. The Triple A of the synthetic CDO presents a different picture. Forty percent or E10m of this tranche would not be deemed Triple A using a typical structure.

Almost all of the Triple A tranches of static synthetic arbitrage deals initially brought to market traded close to levels of cash Triple A tranches. None of them had added structural protections to benefit the senior noteholders. This is still the case with many of the deals brought to market today. The excess spread generated by the cheap super senior tranche was siphoned off for the benefit of the bank arranger, and often most of the excess spread was not even passed through to the equity investor, who sometimes accepted a maximum fixed coupon for the equity investment. The equity coupons looked good relative to a cash arbitrage CDO, but the clear benefit was to the bank arranger.

It is darkly amusing just to combine the words first-loss and Triple A tranche into a phrase that accurately describes a tranche of a CDO. Many investors who bought Triple A tranche synthetic structures didn’t get the best deal possible, because they didn’t know what to ask for at the time. Some day the rating agencies may catch up, but in the meantime, there are several remedies. For instance, once losses exceed a certain threshold, excess spread can be trapped in a reserve account and used for the benefit of the Triple A and super senior noteholders, if required. Investors can also ask for additional spread to compensate for the additional risk.

Uniquely with synthetic deals, investors must also negotiate the language of the credit default swap to make it as comparable to cash assets as possible. All investors in synthetic CDOs should make sure the language used in the credit derivatives contracts does not disadvantage them. This means asking for the cleanest possible language. For instance, for a deal referencing corporate credits, event triggers should be only bankruptcy, failure to pay, and modified restructuring, or better yet should not feature restructuring as a credit event at all. Investors should consider asking for additional compensation to accept additional triggers such as full restructuring. They should also ask for additional compensation if the language does not limit the possibility of delivering discounted assets in the event of default. In other words, investors should guard against being on the wrong end of a ‘cheapest to deliver’ option, as has happened in many European deals in particular.

Every risk has a price, and armed with these caveats, investors can make sure the price is right.

“Adapted with the permission of the publisher John Wiley & Sons, Inc. from Collateralized Debt Obligations and Structured Finance. Copyright ©2003 by Janet Tavakoli.”

JT Note: Subsequent to this article, the rating agencies revised their metholdologies. The situation became much worse.