Monday, June 25, 2007

More on the CDO Put

I got a couple of great questions in the comments section on the CDO Put and since I'm a lazy blogger, I'll just make a new post out of them.

1) "... isn't it the same as saying that if a stock goes down, then everything else being equal, purchasing this stock at a lower price becomes more profitable, so money will flow to this stock and therefore there is a limit how far it can go down?"

No, because I'm speaking of new CDOs being created, not the price of existing CDO tranches. Also because the CDO Put doesn't really limit downside, merely create strong resistance at a certain point. See below.

2) "the cash flow is unchanged, but the present value of the cash flow (probably) changes (if you discount it at a rate including an appropriate risk spread). Couldn't you say the same thing about a traditional bond -- that the cash flow is unchanged when interest rates rise?"

In a way, yes. But a traditional bond is a single credit, or put more technically, its idiosyncratic risk. A CDO is a portfolio of credits, thus systemic risk. Idiosyncratic risk can rise tremendously, but systemic risk can logically only rise so much.

3) (This is the important one). "If spreads widen, it’s presumably because the price of risk is going up. If the price of risk is going up, then the required return on the very high risk, highly levered equity tranches should go way, way, way up. So the fact that the actual return goes way, way, way up doesn’t necessarily make it a better deal. Quantitatively, it may turn out to be true, given reasonable risk-aversion parameters, etc., that it does become a better deal, but proving that would require a lot of messy math (and a bit of economic theory)."

This is right, and the consequences of what my economist friend is saying is that the CDO Put will only hold so long as it is viewed that default risk is relatively contained.

Let's say that credit spreads generally move 30bps wider, and that improves CDO equity returns by 250bps. The extra 250bps is attractive enough to sell the equity, and certainly CDO managers and IB's are happy to make fees, so the CDO engine revs up and creates new demand for credit, thus preventing spreads from drifting wider.

However, let's say that credit spreads are 300bps wider because there is widespread fear of deflation and accounting fraud (see 2001-2002). CDO equity can't get cheap enough and the CDO engine goes into park.

Witness sub-prime related deals today. According to Merrill Lynch the following are "base case" CDO equity return. This is not the actual return, but more the hypothetical return based on where collateral spreads and debt tranches are. I watch this closely and have come to think of it as the "advertised" CDO equity return. In other words, that's the return a generic CDO "would produce" if defaults come in at a historical average level, and hence indicative of what model returns CDOs are being marketed with.

For Mezz Structured Finance (which are mostly residential MBS), the 2006 average was 11.6%. Today it is 65.4%. Let me say that again.

Today it is 65.4%.

This number is so ridiculous as to be meaningless. It is telling you that one could create a new CDO of sub-prime MBS and if defaults were historically normal and if recovery was historically normal, the equity would return over 65%.

By demanding such an outrageously high base-case return, the market is saying they expect defaults and recovery to be substantially worse than historical norm.

Does this disprove the CDO put? No, it just means there are limits. Once the market is tossing out historical norms for defaults and recovery, the CDO put is no more. That's what's happened in the sub-prime market.

But nothing of the like is happening in the loan market. That's because the broad market doesn't expect a wide contagion from sub-prime. So although credit spreads have widened modestly since February, CDO spreads, from AAA to equity are unchanged.

The world of awash in liquidity, and the CDO market is part of that. As long as there are buyers of the AAA and equity tranches, the CDO market will continue to support credit spreads.


Anonymous said...

Is this a vintage issue? 2006 subprime mortgage paper is the worst performing in history due to the lax underwriting standards used and the concurrent housing slowdown.

Perhaps new CDOs can avoid 2006/2007 MBSs and use older MBSs that are performing better as collateral. Or the new CDOs can choose a different asset class.

CDOs used to buy high-yield corporate bonds but around 2002, in light of high defaults, dumped them as collateral in favor of higher-performing leveraged commercial loans.

The other point is that the equity tranches and BBB tranches of CDOs are designed to take significant losses when defaults are unusually high. Why is the media in a lather when this actually happens? The issue is the AA and AAA tranches , which represent 80%+ of CDO deals. If those tranches take losses, then the CDO was not properly designed and the media's panic act will be justified. However, the media doesn't seem to distinguish between the different tranches when talking about downgrades and losses. Not to mention that they focus on market value declines when CDO losses (as opposed to margin call shortfalls) are entirely cash-flow based.

Anonymous said...

In my opinion, the other important point that doesn't get mentioned much in the numerous media reports on CDO's is that the investors in this product are sophisticated institutional investors. This product isn't sold to mom & pop on main street. If the sophisticated institutional investor isn't doing due diligence and research on the product and CDO manager he is buying, then he is an idiot.

Accrued Interest said...

I think it is a vintage issue. Exactly as you say, 1999-2000 vintage high-yield CBOs turned out to be complete shit. By 2001 spreads had blown through CDO Put territory.

Not only is it the lower-rated stuff that's getting hit, but generally you have to be a QIB, which means you have to be worth $100 mil or be an investment pool of that size. If you've got that kind of money to throw around and you don't know what you're doing, I don't feel too sorry for you.

Anonymous said...

Most of the subprime collateral is in Mezzanine CDOs of ABS, which largely invest in the risky BBB tranches of MBS.

Unfortunately, the Mezzanine CDOs of ABS new issue market boomed in 2005 from $60 billion in volume to $200 billion in 2006. Keep in mind that global high-yield issuance in 2006 was only $152 billion in comparison.

Exposure to subprime RMBS among these Mezzanine CDOs? In 2006, it was over 70%!

(See S&P's March 2007 report titled The Subprime Market: Housing and Debt)

These BBB and BB tranches of subprime bonds that are serving as collateral for Mezzanine CDOs only have 8%-10% credit support. The subprime mortgage pools are going to experience losses that will seriously test those credit support levels. Plenty of BBB tranches of subprime bonds will default.

With defaulting collateral, any tranche of the CDO with that underlying collateral could be wiped out. Even a AAA tranche requires cash-flow producing collateral. They can withstand 20% losses but not an utter cash flow drought.

And yet such a drought is possible depending on how the BBB tranche of subprime bonds perform.

Finally, remember that most of the ARM resets have yet to occur. The worst is yet to come. The prospects for these BBB tranches of subprime bonds is precarious at best but all the investors in 2006 Mezzanine CDOs ($200 volume outstanding) have bet on those tranches.

Stay tuned!