Critique of Impure Reason: CDO Anyone? I'll Take Two, Please!

In a perfect world CDOs would have been the perfect investment: for the same level of risk (as distinguished purely by rating, of course) you get more reward, by way of the additional spread or yield on your investment.

Where certain investors -- including pension funds from the US to Europe to Australia -- saw the word "ARBITRAGE" flash in bold red across their screens and quickly rushed to beat the market, others lingered and assessed the risks more closely.

Wisely and slow. They stumble that run fast.

In sum, some investors looked around to find the highest yielding asset that fits their fund's rating criterion (and so they'll necessarily outperform the market, at least for the short-term, and perhaps secure themselves a pat on the back and an increased bonus). Others asked themselves: "Well, even if I can come to terms with the rating agencies' approach to rating these assets -- that the have accurately measured default probability or expected loss, as the case may be -- what else am I paying for, say, relative to a corporate bond? And is it worth it?"

CDO Risks

Without further ado, here's what they're paying for, in terms of risks and expenses, over-and-above the vanilla corporate bond alternative:

• liquidity, or illiquidity risk;
• model risk and/or complexity risk;
• spread or yield risk (somewhat similar to that of a corporate bond);
• interest rate risk;
• payment timing risk (prepayment, extension risk, lumpy payment risk);
• tranche pricing volatility;
• rating methodology changes (given the "newness" of the asset class);
• a slew of operational costs (to monitor/model/mark your position); and
• a host of accounting risks and legal risks that we'll leave for another day

Additionally, certain CDOs brought with them foreign exchange risk, certain basis risks and portfolio market value risks. Leveraged accounts, buying CDOs, took the burdensome funding risk and rating downgrade risk. (This is all not to mention the risk that the rating agencies may be wrong in their assessments of default probability and loss given default.)

While you can't necessarily model the "cost" of each of these risks, and what you should be paid for absorbing each, individually, they are nevertheless quite tangible risks. Especially now.

Ideally, perhaps, a sophisticated investor with strong analytical capabilities might model the deal, run various sensitivity scenarios based on her internal opinion, interview the manager -- and decide that upon sufficient review of the documents and the various "dangers" they allow for, she is is willing to take the risks involved in the investment relative to the yield it provides. It would be suboptimal to say, well, "this looks good to me and geez it pays a whole lot of yield at that rating level, so let's buy it, close our eyes, and hope for the best."

Why? Well essentially you're being paid a hefty reward for being right in your modeling assessment of the deal and in your determination that you can, as a fund, stomach the collective risks of the instrument. If you don't perform the necessary credit analysis, you're not earning your wage: you're rolling the dice or guessing. Tough times, and the volatility they bring, have a sneaky way of separating the guessers from those who performed a thorough analysis, bringing to the fore any portfolio-level risk or support weaknesses, and thereby magnifying certain of the risks of OTC instruments:

• increased illiquidity particularly hurts those buyers who are holding their securities as available for sale (AFS);
• low interest rates decrease payments derived from bonds with LIBOR-based coupons; and
• payment uncertainties, model complexities, and accounting and legal risks and costs accentuate the imbalance between buyers and sellers, driving prices further down. (To compound the matter, lawsuits may discourage both the sale and the purchase of complex securities.)

"Unmodeled," almost qualitative, risks proliferate

- S&P has corrected its ratings on, among others, Founder Grove CLO Ltd., Archstone Synthetic CDO II SPC, and WISE 2006-1 PLC's CDO Notes. Their reason: modeling errors

- S&P has corrected its ratings on, among others, Lorally CDO 2006-2, Tranched Investment-Grade Enhanced Return Securities 2004-11, Longshore CDO Funding 2007-3 Ltd., and Quartz CDO (Ireland) PLC. Their reason: administrative errors

- Moody's and S&P have both confessed to finding errors in their CPDO models

Resources:

1. Arturo Cifuentes on modeling complexity
2. Professor Joseph Mason on the difference in default likelihood between CDOs and similarly-rated corporate bonds
3. On the level of analysis performed by pension funds and endownments prior to purchasing CDOs

(P.S. Feel free to share in the comments section below if you feel we've failed to mention any major risks.)

The KYSS Principle

We're coining a new phrase on this bright-n-sunny summer morning:

KYSS - Know Your Super Senior

Or Know Your Senior Secured. Either way, knowing who's above you in the capital structure can be immensely useful, particularly in the world of defaults.

We've been seeing this in ABS CDO space for a while now, as the contrasting interests and demands of the controlling class holders have determined whether defaulting CDOs were liquidated or accelerated.

And we recently spoke about this in the leveraged loan world too (click here for the full transcript) where banks and CLO managers have possibly different agendas in the corporate loan amendment process:

The big difference is when banks are the lenders the relationship between the borrower and the lender often goes back many, many years and may include businesses like bond underwriting and cash management and other types of solutions that the banks offer. And so the banks are going to be even more incentivized than usual to grant covenant relief and find a solution that allows for continued revenue generation. With institutional investors, on the other hand—and we’re talking CLOs, prime rate mutual funds and the like— they’re “transactional lenders.” In other words, their relationship doesn’t go any further back than the individual loan in question here and so their incentives will be naturally more immediately self‐serving.

Thus, as a prospective subordinated bondholder (i.e., purchaser) it might be wise to find out who is holding the senior secured loan. What would the amendment process look like? What sort of acquisition restrictions will be imposed on the borrower post amendment, and amendment fees and charges will leak to the senior lender. (See for example Richard Kellerhals' recent piece Investors Fume as Banks “Extend and Pretend.”)

So go ahead and KYSS - it may even change the way you see the bond you currently hold.

Impact of amendments on CLOs

Your good friend GP was recently interviewed on the state of corporate loans and CLOs amid the flurry of credit agreement amendments.

You can download the full podcast here.

Collateral Managers and Takeover Opportunities

Collateral managers banding together...here's a link to the report we sent out last week. Enjoy.

Collateral Managers and Takeover Opportunities

Is your CDO Leaking

CDO noteholders, pay close attention to your monthly trustee reports. These are complicated deals and trustees make mistakes. Most of the time, these mistakes cost you nothing but every now and then they'll cost you millions.

For example, a mistake we picked up today revolves around an incorrect implementation of the “CCC Haircut Amount.” It's responsible for leaking approximately $4 million to equity when that amount should have been used to pay down senior notes.

The details...

Here's the definition from the O.M :

Click to enlarge

In simpler words, if this CDO has too many poorly rated assets then it has to carry a portion of these (“The Excess”) at market value (vs. par) when computing the numerator of this CDO’s overcollateralization tests. (The ensuing lower numerator increases the likelihood of an overcollateralization test trip. If such a trip occurs, cashflows that would have otherwise gone to subordinated tranches are redirected to pay down more senior tranches.)

Additionally, the definition specifies that The Excess should consist of those poorly rated assets with the lowest market values (this is typical) but the trustee made a mistake and picked the ones with the highest…

WHAT HAPPENED / WHAT SHOULD HAVE HAPPENED?

Click to enlarge

This misapplication of the CCC Haircut Amount definition causes the class D overcollateralization test to pass when it should in fact fail. Because this test passes, cashflows are leaked out of the deal to equity when they should have been used to pay down the senior tranches in an effort to cure the failing test.

While this is the first distribution date during which the impact of this mistake is felt, chances are that the trustee will keep making it moving forward, ultimately sticking millions in losses to the wrong group of noteholders.

Here are the details of the CCC Haircut Amount calc.:

Click to enlarge

We’ve got a handful of these examples; we’ll try writing more of these in the future if there is an interest…

Assumptions Assumptions Assumptions

"During the recent foolish-extension-of-credit period, I think there was altogether too much reliance on black boxes. Something that comes out of the computer looks quite official; it looks quite precise down to all these little digits. But the fact is, any kind of computer-based model inherently has as its basic assumption that tomorrow will look quite a lot like yesterday. The unfortunate truth is that when you get to a major inflection point, it's precisely because tomorrow does not look very much yesterday." - Wilbur Ross

The rating agencies, among other market participants, have to walk a fine line between maintaining their "long-term" views on long-term securities and being overly adaptive to changing market conditions.

I certainly don't envy them their position: A false step in either direction, and they'll be criticized from Wall Street to Washington.

Before we investigate this double-edged sword, let's consider the original premise or thought process or unspoken truth at the time of the original rating of, say, a CDO tranche. It goes a little something like this:

• this rating is a long-term rating
• given the lengthy maturity (usually more than 10 years away) of the asset, we expect it will go through different economic cycles and so our assumptions should speak neither to the peaks nor the troughs, but to the averages (based on historical data) with some volatility -- i.e., stresses -- built into our assumptions
• as long as the manager behaves as she should relative to the constrictions of the indenture, and as long as the portfolio collateral quality remains within the bounds described, we shall not downgrade you!

Now we return to the question of changing assumptions. As you can imagine, any change in assumptions may precipitate a change in ratings, and so ought to be accompanied by transparency describing the methodological change. A change in rating affects, among other things, the regulatory capital that the holder needs to post against the rated asset and the ability of certain funds to continue to hold the asset. In other words, downgrades precipitate deleveraging. And supply. And price. And therefore recovery. And I could go on and on but this circle is vicious.

From their April 23 press release:

S&P: Criteria Changes And Stressed Collateral Performance Affect TruPS CDO Ratings

We have published several revisions to our ratings criteria for TruPS since the July 2008 trust preferred CDO CreditWatch placements as a result of our observations regarding performance trends and worsening economic conditions, and our view regarding the effect those conditions might have on the performance of TruPS CDOs:

-- "Criteria: Revised Correlation Assumptions For Rtng CDO/CDS Exposed To Financial Intermediaries" published Oct. 3, 2008; this modified the correlation assumptions used for financial institutions held within or referenced by CDO transactions, including bank TruPS CDOs.

-- "Criteria: Correlation Assumptions Revised For Rating Global CDOs/CDS Exposed To Insurance Cos.," published Nov. 6, 2008; this modified the correlation assumptions used for insurance companies held within or referenced by CDO transactions, including insurance TruPS CDOs.

-- "Criteria: Prob Of Default, Correlation Assumps Revise For Glbl CDOs/CDS Exposed To REITs/REOCs," published Nov. 6, 2008; this modified the default and correlation assumptions used by CDO Evaluator for REITs and real estate operating companies (REOCs) held within or referenced by CDO transactions, including REIT TruPS CDOs.

-- "Global Methodology For Rating Trust Preferred/Hybrid Securities Revised," published Nov. 21, 2008; this modified the assumptions Standard & Poor's uses when rating TruPS CDOs generally.

-- "Assumptions: Standard & Poor's Reclassifies Insurance Companies That Issue Debt Securities Owned Or Referenced By Rated CDOs And CDS," published Dec. 23, 2008; this modified the industry classifications used in CDO Evaluator for insurance companies held within or referenced by CDO transactions, including insurance TruPS CDOs.

Stepping back, we're seeing at least five assumption revisions since October 2008. Is this too much? Too little?

Back on April 14, on being downgraded yet again by Moody's, Ambac Assurance responded as follows:
- "While Ambac believes that Moody's is entitled to its opinion of Ambac's financial strength, it notes that this is the tenth such opinion change since January 2008."

As we've described with the current regulation environment, in Hegelian fashion, one tends to over-regulate as a means of "compensating" for under-regulation. Each can be harmful, and hitting the sweet middle-ground is the key. Here we're seeing the responsiveness to severe criticism relating to maintaining static assumptions in a changing environment. The response, naturally, is to proactively rate.

Damned if you do, damned if you don't.

Distress Testing

The Economist put out a piece on the psychology of trading in stressed environments, such as those facing floor traders.

The piece brings to the fore the idea that, in deciding between a low-probability major loss and a high-probability minor loss, the stressed conditions encouraged participants to roll the dice with the major loss.

This theory -- which culminates in traders taking profits too soon and being unwilling to realize losses while they're still manageable -- is consistent with the "loss-aversion winning over utility theory" pieces that stock-trading-psychologist-guru Phil Pearlman writes (see here and here for example).

The nuts and bolts: if a trader has a 100% probability of winning 1 unit, versus a 60% probability of winning 2 units, the theory suggests that the trader often chooses the former option, against the principle of utility theory (since 60% x 2 units = 1.2 units, which is greater than 1).

The alternative is, however, much more troubling especially as it relates to pension funds and government intervention implementation: the willingness to roll the dice and risk a major problem, rather than suffer a sure, minor blow now. On the government level, this "theory" may manifest in an unwillingness to cut rates or even to draw down on the credit line available from the IMF. The United States was a front-runner in cutting rates in early H2, 2007, but some even criticize the U.S for cutting too slowly, too late, with the downturn having begun in 2006. A more rash action may have qualmed fears sooner, and nipped the problem -- now massive -- in the bud.

This trend remains "watchworthy" as companies like Ford reap the rewards of raising capital sooner rather than later and the Japanese banks continue to resist their governments attempt to inject capital, despite their relatively massive exposure to the stock market. Post the G20 meeting, it will be interesting to see how the Balkan (and particularly Baltic) countries differ in their approach towards relying on the IMF.

While it may be acceptable for smaller hedge funds to play ball on the downside, it's incrementally detrimental if systemic-risk-issue-companies, and governments themselves, don't carefully avert losses on the downside.

UPDATE April 22 (Bloomberg): Fitch says Japanese banks may need, yet avoid, public funds

Japan’s biggest banks may need to accept funds from the government as bad debts increase and investors demand higher capital ratios, according to Fitch
Ratings Ltd.

“Capital pressures are growing,” David Marshall , a managing director at Fitch in Hong Kong, said in a Bloomberg Television interview today. Capital weakness and loan losses “might even pressure some of the bigger Japanese banks eventually to have to turn to the government,” he added. “That’s something they’ll resist as long as they can to avoid that stigma.”

UPDATE April 23: Despite wider-than-estimated fourth-quarter loss -- as bad loans spiraled and the global financial crisis cut the value of its investments -- Japan’s second-largest bank Mizuho Financial Group Inc. did not announce any plans to raise money.

From Lemmings to Lemons

"The market-sensitive risk models used by thousands of market participants work on the assumption that each user is the only person using them." - Avinash Persaud, April 2008.

This quote came to my attention via Felix Salmon's Market Movers via Reuters, and it encouraged me to develop our thought process from an earlier piece we put out, entitled Static Measures for a Dynamic Environment.
The point: in a changing environment, one has to proactively adapt modeling assumptions (such as recovery rates and correlations) to reflect those changes.

As Operation Securitization got underway, escalated and then came to an abrupt, sudden halt, each input into the model needed to have been updated due to the gargantuan size of the market -- and its subsequent influence and impact on trading levels -- and the systemic risk is brings with it. For example, the growth of the collateralized loan obligation market (CLOs) from 2001 through 2006 continued hand-in-hand with the growth of the leveraged loan market. With CLOs constituting the majority of demand for these (typically broadly-syndicated) bank loans (roughly 60-65%) the demand base grew in tandem with the supply source. But we saw no adjustment in either recovery rate assumptions (for loans or CLO-issued notes) or in correlation (between loans and CLOs or between loans or between CLO tranches) on the basis of, or necessitated by, this dual, dependent growth.

Surely if the CLOs stop buying, with the demand source halted, loan recovery rates must plunge downwards. And that's what's happened. Indeed performing leveraged loans have recently oscillated between trading levels of 50% and 65%, well below historically realized recovery rate levels for defaulted corporate loans! (70-80%)

We've described this phenomenon in more detail in The Corporate Loan Conundrum. Also, The Elephant in the Room describes our astonishment that certain recovery rate estimates to this day remain unchanged.

The system-wide (systemic) mass-production of securitized tranches helped undermine the value of each in the crisis. The greater the supply, the lower the recovery when things don't work out, and the more correlated they become. And so the banks -- the lemmings -- acting in unison for the most part, created lemons (there are notable exceptions who are still around).

Separately, while my "lemons" are securitized tranches, Brad Setser took the initiative back in 2007 of Turning lemons into lemonade. His lemons are different: they are mortgages; his lemonade being securitized notes.

His article is thought-provoking for many reasons. Here are two: (1) it brings to the fore the economic principle of lemons (think second-hand cars), a principle which relates nicely to the government's purchasing of "toxic assets," and (2) it reminds us of the correlation question: increased correlation improves the quality lower tranches. Why, then, in this market of increased correlation, are the lower tranches of securitized notes not being upgraded? Well, it's a loss-loss scenario for them: correlation, like volatility, increases precisely in the tough times, during which defaults are high. During these times the lower tranches die a quick or slow death in any event, depending on the deal. Superfluous then?

Damaged Goods

With GM in the news on a daily basis, I found myself considering the burdensome scenario of leasing a car expecting to be able to resell it at $x when returned (at the end of the lease period), only to find out it can only be sold at $0.5x.

Well, this possibility (or, now, eventuality) -- not unique to cars -- got me thinking about key inputs to modeling resale value. Assuming these cars are not vintage sports cars, one has to assume they depreciate over time. If they're new cars, they likely depreciate the moment you sign on the dotted line.

My natural assumption would be that new cars would depreciate faster than used cars, from a higher base price, and at a steeper rate (duration and convexity). In other words, if the car acts as collateral for the auto loan, one would assume one would achieve lower loan rates on used vehicles.

To cut a long story short, apparently I was wrong. These auto rates for 48 month car loans available in San Diego are courtesy of Bankrate.com.


P.S. One other consideration may be the credit quality or behavioral patterns of people who buy new cars versus those who buy used cars. One thought, though, is that in this economy the used car purchaser may be the more conscientious. Having said that, I've ignored the possibility that returned used cars may have seen their day (tend to zero value, quickly) whereas some value may remain in returned used cars.

P.P.S. I'm noticing that US Bank agrees with me (re new vs. used car loan rates), but aren't providing this service in San Diego.

"Moody's, we have a problem!"

Last week Moody's announced that it has downgraded and left under review for possible further downgrade its ratings of all classes of notes issued by SVG Diamond Private Equity I and SVG Diamond Private Equity II - deals they should never have rated and should stop rating immediately. (But that's just my opinion.)

Why?

The two deals comprise approximately $717 mm in total. They are CDO transactions, each "referencing a portfolio of shares of private equity funds."

Herein we have problem numero uno: the inherent, obvious lack of diversification by industry. Unlike other CDOs where the rating agencies thought there was diversification, here there never was any such supposition - the underlying are [all or primarily] private equity fund shares (at least according to Moody's press release).

It was diversification and subordination that allowed one to take junk -- it's a technical term, not scathing -- and create investment-grade, or even a AAA. Well, since there's limited diversification here, the junk all acts in tandem, and so the resulting necessity, as we saw above, to downgrade all tranches.

(To be entirely transparent, certain other CDOs, called trust-preferred CDOs or TruPS CDOs, were issued backed wholly by bank or insurance-issued trust preferred securities. These are similarly desirous of diversification, but were rated equipped with the knowledge that banks and insurance companies are heavily regulated and so a lesser default risk. Private equity firms, to say the least, are not heavily regulated.)

From Moody's press release:

SVG Diamond Private Equity is a bankruptcy remote special purpose company incorporated with limited liability in Ireland for the sole purpose of acquiring its interest in the portfolio and certain other assets securing the notes, and issuing the notes.

...

Today's rating actions are primarily a result of the deterioration of the performance of the private equity asset class and the amount of unfunded commitments. The presence of a liquidity facility renders the risk of a default on an interest payment remote.

And now for the juicy stuff:

In reaching its rating decisions, Moody's considered the following important factors:

(1) Cash
The current amount of cash in the structure has been compared to the initial projections. Distributions and drawn-downs have been projected until maturity.

(2) NAV
The Net Asset Value (NAV) of the portfolio as reported by the manager has been compared to the initial projections. Based on the fair value accounting (FASB 157), the NAV presents an aggregated performance metrics.

(3) Public Information Regarding Private Equity
Private-equity funds typically disclose a limited amount of information to the parties involved. Moody's examined the recent disclosure of large publicly quoted buy-out funds with regards to the mark-downs of outstanding Leveraged Buy Outs (LBOs) and Venture Capital (VC) portfolios and used them as guidelines to understand the state of the private equity industry.

(4) Public Equity
Unlike Private Equity, for which historical performance data is available only for the latest 25 years, Public Equity indices provide performance information over a much longer period. As an example, the LPX 50, an index of 50 major publicly traded Private Equity companies, is approximately 67% down since September 2004 (deal inception) and 82% down from its peak in May 2007. In the absence of transparent Private Equity performance data over this time period we can look to Public Equity as a proxy.

(5) Manager's View
The transaction manager has been asked to provide the projected levels of distribution, draw-downs and NAV. Moody's believes that the manager is in a unique position to time the various material elements that impact the cash flow. Moody's applied stressed to the projected levels from managers.

(6) Liquidity Position
By nature, private-equity investors commit capital that will be drawn in the future. Historically, the full amount of committed capital has not been drawn by the Private Equity funds. Moody's believes that the likelihood of a commitment to be drawn during a systemic credit crisis is high.

Moody's has developed a monitoring model for this type of transaction. The model first estimates the cash available over the life of the deal. It also models the liquidity facility dynamically. Based on the factors listed above, Moody's defined three states of future portfolio performance: optimistic, baseline and pessimistic and assessed the probability of losses for each tranche in each of the three states. Haircuts on the projected distributions are in the range between 0% and 40%, depending on the state of the portfolio.

Now we introduce Problems 2 through n:

(2) We're certainly NOT seeing any measure above that takes into account Moody's INDEPENDENT opinion. All we see is "Moody's applied stresse[s] to the projected levels from managers." Sadly, that doesn't take much insight. Nor much investigative research.

(3) These ratings are based on limited historical data: as opposed to the original collateralized bond obligations (CBOs) which were at least supported by corporate bond default rates since at least the early eighties, here Moody's isn't falling back on any substantial historical data.

(4) The limited data they are falling back on isn't even necessarily private equity data: it's public equity and venture capital-type data.

(5) Moody's is relying heavily on the manager's view (!) and projections. Need we say more? This is not entirely dissimilar to a hedge fund investor running in blind despite access to data! Sherlock Holmes would turn in his grave.

(6) Moody's is relying heavily on the manager's NAV. Okay. But how are they combatting potential mismarkings, especially for NAV-lites? I understand if they can't be expected to spot Ponzi schemes, but some cushion on the NAV interpretation would be swell. "The Net Asset Value (NAV) of the portfolio as reported by the manager has been compared to the initial projections." This unfortunately doesn't seem too useful. More useful would be to analyze the NAVs, especially as we're in an economic environment marked by its unwillingness and inability to evaluate illiquid assets (private equity investments are an ideal, typicaly, problematic example). In a market swamped with scandal, including accounting and valuation scandal, solely trusting the key "interested" party simply does not, can not suffice. Especially as an NRSRO - a nationally recogized statistical rating agency. With power comes responsibility.

(7) "The current amount of cash in the structure has been compared to the initial projections. Distributions and drawn-downs have been projected until maturity." Again, I have no confidence that this approach is worthwhile. Firstly, given the market changes, one can't possibly expect anything to be similar to initial projections. Secondly, projecting drawn-downs through maturity -- out in 2024 -- seems a gargantuan, purely academic task.

(8) We still have no knowledge as to what changed that suddenly demanded all tranches be simultaneously downgraded. These deals were rated 3 and 5 years ago. Isn't there a steady realization -- as with all other CDOs they rate -- that the lower tranches have become a credit concern, followed by the mezzanine, and then the senior-most tranches? Was Moody's simply asleep at the wheel?

We have no reason to believe, based on this announcement, that Moody's has any competitive edge over Joe the Plumber in evaluating the quality of this CDO. Aside, perhaps, for the knowledge that "Moody's has developed a monitoring model for this type of transaction." Given these deals were rated in 2004 and 2006, this seems an odd, retrospective remark. From a psychological perspective, this comment appeals to me as a demonstration of innocence by one not accused of anything. The assumption is naturally that they have a (hopefully accurate)model, given the complexity of the transaction and that they have imposed ratings on the issued tranches. The confession therefore, that they have a model, has the opposite affect of being reassuring: it seems Moody's recognizes that they're walking an unnatural, uncomfortable path here, tip-toeing like a cat on a hot tin roof.

It's sad that they ever chose to rate these transactions in the first place. (Why not simply turn it down until you have sufficient data to support such an analysis?)

It's sad that despite the scarceness of data supporting their ratings, and their ratings' heavy reliance on unreliable data, they continue to rate them.

It's sad that each deal's investors continue to pay monitoring fees to Moody's for its scarce, unreliable, spontaneous monitoring.