There are two things that investors want to know about a particular investment offering. First, does the offering have a favorable return pro- file? Second, how does the variability of the offering’s return fit in with the preexisting portfolio?
We have gone into detail about how to assess the return profile of CDO equity, making use of current values for many important vari- ables, historical values for collateral defaults and recoveries, and cash flow modeling. It turns out that the answer to the second question above, on return correlation, is based on the nonrecourse leverage nature of CDO equity.
CDO equity is basically a leveraged position in the assets of a CDO, with the CDO’s debt tranches being the financing for the equity posi- tion. CDO equity receives whatever cash flow remains after satisfaction of debt claims. Equity sustains the risk of collateral asset payment delays and credit losses, but also receives the upside if CDO assets gen- erate cash flow in excess of debt tranche requirements. Meanwhile, debt tranche holders only have recourse to the CDO’s assets, and cannot make any additional claims against equity holders. CDO equity holders are not at risk for any more than their initial investment.
CDO equity holders receive financing that is in place for up to 12, or even 15, years. Moreover, the financing rate is locked in, either at a fixed rate or more commonly at a spread above a designated floating index (usu- ally LIBOR). CDO debt is subject to early amortization only if asset quality deteriorates according to objective measures (such as overcollateralization tests). In that case, principal repayment is due only to the extent the asset portfolio provides cash flow. Asset sales are never required.
The features of financing available to CDO equity are in contrast to the repo market or other short-term secured financing arrangements. In those, financing rates can fluctuate and higher levels of security (larger collateral haircuts) can be demanded. In fact, financing can be pulled altogether with little warning. Collateral assets are subject to sale by the creditor, and creditors have recourse to the borrower if the collateral is insufficient to extinguish the debt. Leveraged investors in a CDO’s assets (also known as. CDO equity holders) avoid all these risks.
Return Correlation Simplified
In Exhibit 19.12 we graph CDO underlying returns and CDO equity returns under the simplifying assumptions that (1) CDO equity does not receive any cash flow until the claims of CDO debt holder are com- pletely satisfied; and (2) CDO equity holders do not take advantage of
c19-EquityReturns Page 374 Monday, March 6, 2006 11:26 AM
certain optionalities inherent in their position. We will relax these assumptions and explain their significance shortly.
The graph’s horizontal axis depicts returns on the CDO’s underlying assets. In our example, the best case is that the CDO portfolio suffers no defaults and returns its promised yield of 10%. In the worst case, all the assets in the CDO’s portfolio default without any recovery and the port- folio’s return is –100%. The thin line going up at a 30-degree angle rep- resents the tautology that the return of the underlying CDO assets depends on the return of the underlying CDO assets. Therefore, it also goes from –100% to +10%.
The thick line in Exhibit 19.12, which moves along the bottom of the graph and becomes a rising dotted line at the right of the exhibit, represents CDO equity returns. We assume that equity contributes 10%
of the funding for the CDO’s asset portfolio and that 90% of the CDO portfolio is funded with CDO debt costing an average of 5%.
There are two distinct regions in Exhibit 19.12 with respect to the return correlation of CDO equity and the CDO’s underlying assets. In the right side of the exhibit, when CDO equity return varies between –100%
and +55%, there is a strict mathematical relationship based on the excess cash flow from the underlying CDO collateral over that which is necessary to service the CDO’s debt. All of this excess goes to CDO equity and every extra dollar of excess cash flow is an extra dollar of
EXHIBIT 19.12 Two Regions of CDO Equity Return Correlation
376 CDO EQUITY
return to CDO equity. In this region, at the right of the exhibit, the return correlation between CDO assets and CDO equity is a perfect 1.0.
Between CDO underlying returns of –100% and –5%, CDO equity’s return is flatlined at a –100% loss in our example. This is because, as we have said, of the nonrecourse nature of CDO equity. As equity holders in a corporate entity, CDO equity holders are not responsible for the debts of the CDO and can only lose the purchase price of their equity investment. In this region, the return correlation between CDO assets and CDO equity is zero.
The simple, and pretty good, answer to the correlation of CDO equity returns and the CDO asset portfolio is thus: when CDO equity is not completely wiped out, the return correlation is 1.0; and when CDO equity is wiped out, the return correlation is zero. These correlations are based on the cash flows of the CDO equity and CDO assets over their lives, as opposed to monthly price movements. We have argued why this is the only reasonable way to look at CDO equity returns.
Having correlated CDO equity returns to the returns of the underly- ing CDO assets, it is simple to compare the return correlation of the underlying CDO assets to what might be in the investor’s portfolio through the traditional practice of using monthly returns, as we show in Exhibit 19.13. So, if CLO equity is firmly in-the-money, its return corre- lation with the loans in its portfolio is 1.0. The exhibit shows the monthly return correlation of loans to other assets, which is our proxy for the return correlation of CLO equity with those same assets.
Relaxing the Assumption of Strict Equity Subordination
We mentioned making some simplifying assumptions in the above analysis and now it is appropriate to discuss their implications. Our first assump- tion was that CDO equity holders do not receive cash flow until and unless CDO debt holders are repaid in full. But in any realistic collateral default scenario that produces losses on CDO debt, CDO equity will ini- EXHIBIT 19.13 Monthly Asset Return Correlations
HYL HYB IGB S&P U.S.T
High-yield loans 1.00
High-yield bonds 0.48 1.00
Investment-grade bonds –0.06 0.23 1.00
S&P 500 0.14 0.51 0.02 1.00
U.S. Treasuries –0.11 0.04 0.91 –0.09 1.00
c19-EquityReturns Page 376 Monday, March 6, 2006 11:26 AM
tially receive cash flow. Later, overcollateralization tests become binding and redirect cash flows to paying down debt holders or purchasing new collateral assets. So some cash flow will leak out of the CDO structure to equity holders in scenarios where CDO debt holders are not repaid in full.
How much depends on the timing of collateral losses and the strictness of the CDO structure in cutting off cash flow to equity holders.
What this means with respect to the previous analysis is that we have been too pessimistic about CDO equity returns in cases when CDO debt holders are not repaid in full. In fact, rather than flat-lined at –100% in Exhibit 19.12, there would be some return to CDO equity holders in all but the very worse collateral return scenarios. While relaxing our assumption about strict equity subordination messes up the CDO returns we drew in Exhibit 19.12, it does so in a nice way for CDO equity, as it means the equity will get a higher return than we supposed.
This is particularly true for CDO equity existing on the cusp of having value or being worthless.
How does relaxing the assumption about CDO equity subordina- tion affect the correlation of CDO equity returns to those of other asset returns? The dependence of CDO equity returns to the strictness of the CDO structure makes CDO equity returns less correlated to any asset return. This is because the strictness of a CDO’s structural features is uncorrelated to the returns of financial assets.
Relaxing the Assumption of Nonoptionality
Our second assumption in the above analysis was that CDO equity holders do not take advantage of certain options inherent in their posi- tion. In the previous chapter, we discuss the two different options CDO equity holder’s have on the CDO’s assets:
1. On the market value of assets in the CDO portfolio that is exercised by liquidating and unwinding the CDO.
2. On the after-default cash flow of the CDO asset portfolio that is reaped by simply waiting to see how actual defaults and coverage tests interact to produce equity cash flow.
Market value and after-default cash flow optionalities break up our hold-to-maturity analysis, but again in a positive way for CDO equity.
Our assumption, this time about the CDO equity not being able to take advantage of certain optionalities, underestimates CDO equity returns.
These optionalities also serve to further uncorrelate CDO equity returns to other assets because their existence and their use is uncorrelated with the return of financial assets.
378 CDO EQUITY
CONCLUSION
In this chapter, we explained why research about historic CDO equity returns, particularly with regard to the monthly volatility of equity returns, their Sharpe ratios, and the correlation of CDO equity returns to the returns of other assets, is misguided. We described the flaws of a type of CDO equity analysis built on dubious data and methods. We queried whether, even if return data for old CDO equity were available, how applicable they would be to new CDO equity issues.
We proposed what we believe to be an eminently reasonable, unsur- prising, and even commonplace method of looking at CDO equity:
Model what can be modeled in cash flow analysis, assess the relevance of historic defaults and recoveries, and test their implications in the cash flow model. We provided historic defaults and recoveries for high-yield loans and structured finance assets using rating agency data.
Then, with regard to the correlation of CDO equity returns to CDO underlying asset returns, we offered a novel perspective that takes into account the nonrecourse nature of CDO equity. After relating CDO equity to CDO asset collateral, we used traditional methods (where there are adequate data) to follow the chain of correlations to other assets in the investor’s portfolio. Finally, we relaxed assumptions that we made in the analysis of CDO equity returns and return correlation and showed that these assumptions underestimate CDO equity returns and overestimate CDO equity return correlation.
c19-EquityReturns Page 378 Monday, March 6, 2006 11:26 AM
Eight PART
Other CDO Topics
p08 Page 380 Monday, March 6, 2006 11:26 AM
CHAPTER 20
381
Analytical Challenges in Secondary CDO Market Trading
he secondary CDO market has grown rapidly. Annual par traded totaled approximately $2 billion in 2002, then soared to $14 billion in 2003 (about half of that from Abbey National sales), to $25 billion in 2004, and $30 billion in 2005. More importantly, the secondary market has become an important part of CDO investor thinking, providing a second chance to acquire CDOs and rewarding rigorous analysis.
One reason for the secondary CDO market’s growth is the signifi- cant increase in resources available for analyzing outstanding CDOs, chiefly the availability of CDO-specific cash flow models. However, potential buyers of secondary CDOs still face challenges when consider- ing a purchase. A very liquid secondary CDO market, trading with the ease of corporate bonds, will probably never be achieved. Cash flow CDOs, with idiosyncratic portfolios, management, and structures, will always be complex and take time to analyze.
In this chapter we first review secondary market developments and pitfalls. Next, we provide a step-by-step guide to evaluating a secondary CDO. We pay particular attention to two popular analytical methods:
net asset value (NAV) analysis and cash flow modeling. These two tech- niques have improved investors’ understanding of CDOs and therefore secondary market liquidity. But they must be implemented with care.
We demonstrate that NAV can produce results far from the actual trading level of secondary CDOs. The simple NAV calculation does not address the different cash flow characteristics that CDO tranches can have. In cash flow modeling, the choice of default scenario is vital to the relevance and comparability of results. For structured finance-backed CDOs, we suggest a way to determine default scenarios based on the
T
382 OTHER CDO TOPICS
ratings and prices of underlying collateral. For high-yield, corporate- backed CDOs, we review and formalize market practice.
Finally, we recommend the terms primary market purchasers should demand on their CDOs to ensure that the bonds have the greatest possi- ble liquidity in the secondary market.