As with the buying and selling of any financial instrument, the trading of secondary CDOs is not without its pitfalls. The list of hazards we will discuss includes:
■ Reliance on inaccurate or incomplete information.
■ Over-reliance on net asset value analysis.
■ Ignoring the potential for collateral prepayments and CDO calls.
■ Inaccurate reinvestment assumptions.
■ Ignoring manager behavior.
■ Over-reliance on cash flow models.
Reliance on Inaccurate or Incomplete Information
Perhaps the most basic mistake, in terms of being clearly right or clearly wrong, is reliance on inaccurate or incomplete information about a CDO. There are many factual points that can be got wrong about a CDO, chief among them are its portfolio and the priorities of its cash flow waterfall. Obviously, getting these points right is important to the
1 The “basis” of a bond is its credit default swap premium minus its spread above LIBOR in the cash market. Thus, if credit protection on a senior AAA CDO tranche can be purchased for 10 basis points and that tranche’s spread above LIBOR is 25 basis points, there is a 15 basis point negative basis (10 – 25 = –15). A “negative basis trade” in a CDO context is when, for example, Party A buys a AAA CDO tranche and Party B sells credit protection on the CDO tranche to Party A. The funding of the CDO tranche has therefore been separated from the assumption of its credit risk. Typically, a financial institution with a low cost of funds buys the tranche and a monoline insurance company writes the credit default swap.
c20-AnalyticalSecondaryCDO Page 384 Wednesday, April 19, 2006 12:06 PM
proper valuation of a CDO. We mentioned before the increased avail- ability of CDO information and analysis as a factor in expanding the secondary market, so there is now little excuse for getting the basics of a CDO wrong.
Overreliance on Net Asset Value Analysis
Briefly, because we address this later in detail when we provide a step- by-step guide to evaluating a secondary CDO, the calculation of a CDO tranche’s net asset value begins with the aggregate market value of the CDO’s assets. This figure is adjusted by the market value of interest rate hedges. That usually means subtracting the amount by which an interest rate swap is out-of-the-money to the CDO. Finally, the par amount of CDO tranches ranking senior to the tranche in question is subtracted.
The result, the net asset value of a particular tranche, is divided by the par amount of the tranche and the percentage is considered by some to estimate the tranche’s market value.
NAV analysis says something about the relative quality of assets, in comparison to CDO liabilities, that cannot be evaluated by traditional par value overcollateralization tests. However, the NAV concept is unre- alistic in that the CDO is not going to be liquidated unless all of its debt tranches can be fully repaid. But the big problem with NAV analysis is that it does not take into account the cash flow characteristics of a CDO. The best example is that of a non-PIK CDO tranche—that is, one whose coupon comes before any overcollateralization test. While this tranche may have zero NAV, it most likely has a pretty sure interest stream that may have a significant present value.
Ignoring the Potential for Collateral Prepayments and CDO Calls
In 2004 and 2005, some sellers of CDOs ignored the potential for CDO collateral prepayments and CDO calls at great cost to themselves as high yield bond and loan prepayments were much greater than expected. For example, loan repayments (both scheduled amortization and prepayments) reached a record $99 billion (about half the outstand- ing amount of loans) during one 12-month period in 2004 to 2005.2 For CBOs and CLOs out of their reinvestment period, significant cash has flowed through to repay CDO debt.
In some cases, senior tranches paid down faster than expected, pro- viding windfalls to investors who purchased tranches at a discount. In
2 S&P/LSTA Leverage Loan Index: February 2005 Review, Standard & Poor’s, March 9, 2005.
386 OTHER CDO TOPICS
other cases, PIK-ing tranches returned to cash flow status faster than expected, also increasing their value. Sellers, obviously, suffered oppor- tunity costs in letting their CDOs go too cheaply.
Similar to collateral prepayments are opportunistic calls of CDO debt and unwindings of CDOs. As we mentioned before, this cannot be accomplishment without paying all CDO debt holders off in full.
According to S&P, at least 40 CDO were called between 2003 and 2005, including 12 emerging market CDOs, six high-yield CBOs, two SF CDOs, and nine CLOs. Investors who sold tranches of these CDOs at less than par before they were called also suffered an opportunity cost.
Inaccurate Reinvestment Assumptions
Modeling CDOs at their initial reinvestment assumption with respect to spread is obviously inaccurate in a declining spread environment. In fact, in the aftermath of dramatic spread tightening, assuming that a CDO can reinvest at all has been a bad assumption for many CDOs vio- lating their minimum weighted average spread (WAS) requirement. Built up cash in CDOs that cannot reinvest can usually be passed out of the CDO even before the end of the CDO’s reinvestment period at manager discretion. And cash balances usually have to be passed out of the CDO after the CDO’s reinvestment period.
Ignoring Manager Behavior
CDO investors ignore manager behavior at their peril. While class war- fare may be discredited as the driver of human history, it still is a good model for the way some managers manage their CDOs. Whether a CDO manager is equity friendly or debt friendly determines whether overcol- lateralization tests are gamed or allowed to fail and thus divert cash flows from subordinate tranches to senior tranches.
Overreliance on Cash Flow Models
Finally, as much as we depend upon them, we list over-reliance upon cash flow modeling as our final secondary CDO pitfall. One problem is the way most cash flow modeling is done, via constant annual default rates. The fact is that some structures don’t show their strengths or weaknesses in these rather unrealistic default scenarios. For example, the value of equity caps and turbo amortization only becomes apparent in scenarios where defaults are delayed.
c20-AnalyticalSecondaryCDO Page 386 Wednesday, April 19, 2006 12:06 PM