Deconstructing ABACUS

Goldman’s controversial “ABACUS 2007-AC1” synthetic CDO turns out to be a very complicated deal. This is not your grandfather’s vanilla mezzanine RMBS synthetic CDO. It is, in some sense, a supersynthetic CDO.

There’ve been some excellent posts dissecting the deal, including…

Also, the formal prospectus is now available, as well as a marketing “flipbook“.

In what way was ABACUS a “supersynthetic CDO”? Despite notionally having seven classes of investors, just two classes of notes were actually sold. When I read this at Alea, it blew my mind. The only notes that were sold were AAA debt, from senior (but not “super senior”) tranches. I didn’t understand how this could work. CDOs turn low-class debt into AAA gold by segregating losses. Senior notes are made safe by shifting losses to junior tranches, and remain safe until the junior tranches are wiped out. I had seen synthetic CDOs with unfunded senior classes, in which case the issuer retains some risk if the CDO fails catastrophically. But if there are no junior tranches, who takes the first loss? Who stands in the line of fire to protect AAA noteholders?

I spent some time squinting over the prospectus to understand. But there is no clearly stated explanation. On the contrary, there is a lot of language like this:

On (i) each Payment Date and (ii) any other Business Day on which Currency Adjusted Notional Principal Adjustment Amounts are paid by the Issuer to the Noteholders, the Class SS Notes will be senior in right of payment to the Class A-1 Notes, the Class A-1 Notes will be senior in right of payment to the Class A-2 Notes, the Class A-2 Notes will be senior in right of payment to the Class B Notes, the Class B Notes will be senior in right of payment to the Class C Notes, the Class C Notes will be senior in right of payment to the Class D Notes and the Class D Notes will be senior in right of payment to the Class FL Notes.

That sounds like the standard CDO waterfall. But in reality there was nowhere for the water to fall, because no B, C, D, or FL notes were sold. If losses were allocated to any investor, they would be allocated to AAA tranches. So what was going on?

The ABACUS prospectus doesn’t say. But there is a hint. Rather than buying credit default swaps on the aggregate reference portfolio, then dividing the cash flows among the tranches based on seniority, the CDS payments are calculated separately for each “series” of notes (where the series are subdivided by class). In other words, each class of notes writes its own distinct insurance policy.

As best as I can tell, there are two distinct levels of abstraction in the ABACUS deal. First there is the reference portfolio, a hypothetical portfolio of debt. Then there is a notional CDO, a hypothetical entity that we imagine to have purchased (or synthesized) the reference portfolio. We pretend that this notional CDO is “fully funded”, with a $1100M SS tranche (“super senior”), a $200M Class A-1 tranche, a $280M Class A-2 tranche, a $60M Class B tranche, a $100M Class C tranche, a $60M Class D tranche, and a $200M FL tranche (“first loss”). In reality, no one has purchased any of the reference portfolio, and the notional CDO, which would have required $1.8B $2B of investor interest to build, was never constructed. Instead, the notional CDO forms the basis for a thought experiment: Given any performance scenario for debt in the reference portfolio, we can compute the loss that would have been experienced by holders of the various tranches. So, we could write a kind of swap (somewhat different from an ordinary credit default swap), whereunder a “protection buyer” pays a predetermined, fixed spread and a protection seller pays the losses that a hypothetical holder of a tranche in the notional CDO would have experienced.

Effectively, the seven tranches of the notional CDO serve to define seven new kinds of bets that one could take on the reference portfolio. Since these bets are designed to mimic the experience of investors in a real CDO, S&P and Moody’s were able to associate ratings with these bets. However the bets themselves — highly customized variants credit default swaps — are not securities.

For a regulated entity that wished to hold AAA debt, securities had to be constructed based on these bets. The actual “ABACUS 2007-AC1” legal entity offered synthetic securities designed to mimic the experience of tranches in the notional CDO. It did so in the usual way, just as a commodity ETF or “vanilla” synthetic CDO would: The entity accepts money from investors, and uses those funds to purchase ordinary, low-risk debt. (In this case, the low risk debt was not so ordinary; it was itself a synthetic security. But we’ll set that aside.) The entity then takes a side bet. Investors’ earnings are interest on the low risk debt adjusted by the gains or losses they experience on the side bet. The net effect of all this is that buyers of “notes” from the entity experience outcomes that are almost exactly as if they had invested in a tranche of real CDO. However, notes synthesized this way do not need to be backed by a funded CDO structure (cash or synthetic). In fact, the scheme completely eliminates all constraint on the quantity of funding invested in a given tranche, and severs any relationship between the quantity of funding and the characteristics of the securities. Goldman could have sold $1 worth of Class A-1 notes or a $1T dollars of Class A-1 notes, as long as it was able to make itself comfortable with taking the other side of the Class A-1 side bet. It happened to sell $0 worth of Class C notes, but it could have sold any quantity, without altering the characteristics of the notes or the structure of the notional CDO.

Once you “get it”, the scheme is not very difficult to understand, and it is clever. But it is not clearly described in either the ABACUS pitchbook or prospectus. I don’t know why the three-level structure is not clearly diagrammed (reference portfolio -> notional CDO -> funded entity that replicates the experience of arbitrary tranches in the notional CDO). “Notional CDO”, by the way, is my term. It is nowhere in the prospectus or pitchbook. The distinction between the notional CDO and the actual funded entity are blurred in the documentation. Perhaps the structure of this sort of deal would be obvious to insiders, or perhaps there are clearer descriptions elsewhere, in documents that have yet to be made public. Both Alea and David Harper have pointed out that this structure is similar to a “bespoke” or “single-tranche” CDO. Effectively ABACUS describes a hypothetical cash CDO with seven tranches, then chops it up into seven “single-tranche” CDOs, only three of which were ever actually invested. But in none of the documents is it represented as a bespoke CDO.

A remaining issue that has not received much scrutiny is how the deal was priced. IKB earned LIBOR + 0.85% on its Class A-1 tranche, prior to any credit events. Both ACA and IKB earned LIBOR + 1.10% on Class A-2 notes. In a cash or more vanilla synthetic CDO, the above-LIBOR cash flow to CDO investors is determined by the credit spread on the underlying debt, potentially plus a “basis” if demand for insurance has pushed the market price of protection above the underlying’s credit spread. Effectively, cash flows into the structure are market determined. (The allocation of spread between the tranches is an internal matter among the CDO’s investors.) With ABACUS or a bespoke CDO, there is no market in the tranche-specific credit default swaps and no security with an observable credit spread that can serve as a basis for pricing. So the price of protection must be negotiated between the protection seller (the ABACUS SPV and its investors in this case) and the protection buyer (usually the deal’s sponsor) without a very clear benchmark. Disclosure of the fact that there was an adversarial counterparty on the other side of the deal would likely have affected the character and perhaps the outcome of those negotiations. Since investors may have believed the ABACUS deal was offered and underwritten at Goldman’s initiative, it’s unclear whether there were active negotiations at all, or whether ABACUS investors simply accepted spreads computed by Goldman on the theory that as customers of a reputable bank they would be given reasonable prices. (“Fair” prices would have to be modeled, and modeling a fair price of a bespoke CDO tranche might be within the competence of an investment bank but beyond the competence of even “sophisticated” institutional investors.) Sponsors of bespoke CDOs often hedge their exposure in public markets, so ABACUS investors need not have suspected that there would be an identifiable counterparty, who was also a customer of Goldman’s, negotiating against them on price. Alternatively, Goldman undoubtedly had more efficient means of hedging its exposure that it otherwise would have, since it could just lay off the risk on Paulson. So Goldman might have been able to offer unusually good pricing to ABACUS investors. We cannot say a priori whether ABACUS investors ended up receiving better or worse pricing than they would have had Goldman underwritten this deal on its own initiative and hedged its exposure. But investors did not have the opportunity to negotiate price in full awareness of an adversarial counterparty, so the fairness of the spreads investors received merits further examination.

To summarize, ABACUS defined seven “side bets” based on the performance of the reference portfolio. Under each bet, one party would insure the losses of a hypothetical tranche of a notional CDO in exchange for fixed payments from the other party. The ABACUS legal entity synthesized securities based on two of those side bets, and sold those synthetic securities to IKB ($150M) and ACA ($42M). But as Alea points out, the largest “investment” — by ACA via ABN-AMRO — was not actually a purchase of notes from the ABACUS SPV, but an unfunded side bet. ACA/ABN took a $909M “long” positions in one of the seven side bets, with Paulson (via Goldman Sachs) on the other side. This was an unfunded CDS-like arrangement that occurred some time after the ABACUS legal entity was formed and funded.

I think in judging Goldman Sachs’ behavior, the fact that the ACA/ABN “investment” was a side bet arranged after the deal closed is important. The SEC’s main allegation, that Goldman was less than candid about Paulson’s role during the selection of the reference portfolio, would have affected all parties, IKB, ABN-AMRO, and ACA, both as noteholders and bond insurers (side bettors). But the question that I find most interesting is whether or not Goldman mistreated investors by virtue of a conflict between its roles as market maker and underwriter. That conflict directly affected only IKB and ACA as purchasers of newly underwritten notes. The ACA/ABN “wrap” of the super senior tranche occureed after the ABACUS LLC had been underwritten, so Goldman was only a counterparty to ABN/ACA at that point in time.

Update: Correction: IKB invested the A-1 tranche, not ACA as originally stated. Many thanks to commenter gennitydo for pointing out the error.

Update 3-May-2010: Yves Smith publishes a note from an anonymous correspondent claiming that ABACUS was just a failed underwriting of a vanilla CDO, not several “singe tranche CDOs” as described above. I think her correspondent is mistaken, and stand by the post as written.

If ABACUS had been constructed as a vanilla synthetic CDO, but the junior tranches had been left unfunded, Goldman would have been on the hook for that risk (as well as the risk of the super senior tranche and the unfunded portion of the Class A-1 and A-2 tranches). Goldman would have lost at least $708B on the deal if that had been the case, probably much more, depending on how worthless the super senior tranche turned out to be. Goldman could have synthesized the full reference portfolio and then dynamically hedged its exposure to the whole unfunded portion of the structure, but that would have been an elaborate and inefficient means of reaching an economically identical result. The prospectus notes that the structure would sell CDS by series of note, where series are within-tranche groupings, which it would not have done if it were synthesizing the full reference portfolio. ABACUS was built from single-tranche CDO’s, with Class A-2 notes covering a 21% – 35% slice of a notional CDO built from the reference portfolio and Class A-1 notes covering a 35% – 45% slice, while the unfunded but eventually insured super senior tranche was 45% – 100%. No one funded or ever bore the risk of the 0% – 21% bit.

Many thanks to the indispensable jck of Alea for great comments on an early draft of this post. All the dumb mistakes are mine; the smart stuff is jck’s benign influence.

Update History:

  • 26-April-2010, 6:45 a.m. EDT: Corrected misstatement of which parties invested which tranches, with thanks to commenter gennitydo.
  • 3-May-2010, 3:00 a.m. EDT: Added update re a purported debunking of this description published at Naked Capitalism.
  • 3-May-2010, 3:45 a.m. EDT: While reviewing the piece after its alleged debunking, I notice that I am arithmetically inept. It would have taken $2B, not $1.8B to fully fund the structure. Corrected in the text with the old value scratched.

59 Responses to “Deconstructing ABACUS”

  1. Nemo writes:

    Isn’t what you are describing the very definition of a “synthetic CDO”?

    Buy risk-free debt, sell protection on risky asset via CDS. Earn the risk-free yield plus the risk premium, which equals the yield the risky asset would have provided?

    There is no necessary relationship between the quantity of the underlying “cash” entity in existence and the amount of the “synthetic” entity in existence. So it is not too surprising that the former could be “zero”. Although it is a little strange. Imagine if you could buy and sell options on a company that does not exist, with payoffs based on how the hypothetical company would perform if it did exist…

    Almost all of this stuff sounds to me like it ought to be illegal.

  2. […] – The ’super-synthetic’ CDO. […]

  3. Dukot writes:

    This is exactly how a plain vanialla synthetic CDOs work. The fact that a senior or a junior tranche is unfunded makes no difference to the investor. The investor get paid or loses money when his/her tranche starts loosing money. The investor is agnostic to whether there was a junior tranche or not as it does not affect that person. In fact it is precisely due to this reason, ie. some tranches can be unfunded, that synthetic CDOs are written. You dont need a junior tranche as you are not actually investing money in the underlying bonds, you are just writing a CDS against them.

  4. crookery_ writes:

    Steve, once you’ve got over synthetic = hypothetical then the rest follows as you have now illustrated. Simple analogy is that of an IB brokering, say, 80-50 putspreads on a basket of stocks selected by one party. No one wants, needs or even does ever own the stocks. They are just a reference for an OTC contractual expression of a view. The rest of the structuring (in the case of Abacus) is obfuscation to make the synth have as much as a feel of a cash CDO as possible, as that’s what blind (not blindfolded) investors were used to buying.

  5. As near as I can tell, the point was to get an investment with a desired credit rating. The rating was more important than any product, which is weird, because the rating is supposed to be an evaluation of a product’s risk. This system seems clearly based on the reliability of the ratings, and the fact that people saw credit ratings like the grading of meat, when it’s much more subjective. It looks to me more like a bet against the reliability of the credit ratings than a bubble. The bubble was caused by the poor grading of risk. Without that, it’s hard to see how this thing blows up. To me, the poor grading of risk was likely a product of fraud, negligence, fiduciary mismanagement, and collusion. That’s what Paulson saw. That’s what needed to be disclosed. That’s the meaning of full disclosure. It looks to me like GS is trying to resurrect caveat emptor in its defense. That’s the point of GS noting the sophistication of the investors.

  6. gennitydo writes:

    If the argument is that IKB (which btw had both A-1 for 100 and A-2 for 50) and ACA were mis-treated as investors because GS failed to disclose the role of Paulson, then shouldn’t it be ACA in the dock rather than GS? It was ACA as much as GS that claimed to have “independently” selected the reference portfolio.

    Even if it was not clear what role Paulson had in the transaction, ACA was clearly aware that Paulson had been partly responsible for the so-called independent selection of the reference portfolio. ACA as much as GS passed its work off as independent to IKB. Moreover, by the time the offer document was prepared, ACA must then have known Paulson was not long the FL tranche as the FL tranche was unfunded. ACA must have worked out Paulson’s role at that point for sure.

  7. anon writes:

    “Notional CDO” is an unnecessary concept.

    E.g. super seniors are generally not funded.

  8. Steve Randy Waldman writes:

    Nemo & Dukot & anon — No, this is a very different product than a “vanilla” synthetic CDO, at least as I have learned and use the term. I think my use of terminology is pretty standard, because the CDO industry has a distinct and notably more exotic category for products such as this, the “bespoke”/”single tranche” CDOs.

    Let’s understand the distinction: A vanilla synthetic CDO is a real, waterfall-structured CDO, in which first losses are allocated to junior tranches, protecting more senior tranches. It is distinguished from a cash CDO simply by virtue of its assets. A cash CDO actually owns a portfolio of assets. A synthetic CDO does not actually own the underlying assets, but synthesizes them. If the assets are bonds, the synthetic CDO synthesizes the reference portfolio by purchasing nearly credit-risk-free debt and layering a book of CDS positions on top of that.

    Crucially, the whole reference portfolio is synthesized, and the whole structure is spoken for.

    anon — You are right that often supersenior tranches are not funded, in cash or synthetic CDOs. But they are still spoken for in vanilla structures. The arranger of the CDO retains the risk of the super-senior tranche, and is de facto invested in the CDO.

    The “notional CDO” in this structure is a different ballgame. It is entirely notional. More vanilla tranches can leave senior tranches “unfunded”, but cannot leave equity or junior tranches unfunded. Someone has to bear the first loss. You could not construct a CDO with unfunded junior tranches that owns or synthesizes the whole reference portfolio. If you want to leave junior tranches unfunded, you have to synthesize investments on a tranche-by-tranch basis. And if you are synthesizing tranches, you are in a much more dangerous and exotic place than if you synthesize whole bonds or portfolios of bond.

    Why more dangerous? Because it is always possible to price credit risk on bonds or portfolios of bonds, but it is impossible to get a market price of credit risk for bespoke tranches. Real bonds have observable credit spreads, and perhaps observable CDS prices. The credit risk of bespoke tranches must be modeled, but models rarely agree and can be tweaked. Negotiating a “single tranche CDO”, or several bundled single-tranche CDOs as with ABACUS, is necessarily involves greater diligence or greater trust than negotiating a cash or vanilla synthetic CDO.

    [Note: I’ve edited this comment after posting to strike “cash or”. Economically, the issuer of a cash CDO may hold the supersenior tranche, bearing the risk just as the issuer of a vanilla synthetic would if it left the supersenior unfunded. But in a cash CDO, leverage is explicit, so as a matter of accounting all traches must be funded.]

  9. Steve Randy Waldman writes:

    crookery_ — I’m pretty comfortable with synthetic = hypothetical. What interested me about this structure is that there are two layers of hypothetical: a hypothetical reference portfolio, and the “notional CDO”. “Vanilla” synthetic CDOs just have one level of indirection.

    I like your analogy with synthesizing an option spread — that’s what bespoke CDOs feel like to me. And the product really is very clever.

    I am inclined to agree with you regarding the obfuscation of the exoticness of the structure in the filpbook and prospectus. On the one hand, as a software guy, I’m sympathetic. Core ideas of software architecture include encapsulation and interfaces: As long as one holds the interface and behavior of a module constant, how it is implemented should be regarded as a black box from the outside to give developers maximal freedom to implement and improve the module on the inside. Synthetic CDOs are designed to mimic cash CDOs, an outside investor shouldn’t much care that it’s synthetic. Investors in a single tranche of a CDO should in theory not care whether the CDO is cash, “vanilla” synthetic, or single-tranche. In a perfect world, investors would not need to care how their CDO was implemented: they would just need to know the economic characteristics of the underlying asset portfolio and the attachment points of their investment.

    But in practice, both in software and in finance, the details of implementation sometimes matter, despite attempts to hide and abstract away from them. Unfortunately it does matter whether one holds a cash CDO, vanilla synthetic CDO, or bespoke CDO, and investors should be clearly informed of which they are buying.

    Why? Because even a vanilla synthetic is subject to real-world risks that would not apply to a cash CDO. An synthetic is subject to counterparty risk on the underlying and the collateral securities, as well as credit risk on the synthesized debt. That counterparty risk might be minimal, but it is also irreducible. A synthetic may need to roll or renew its positions, in which case its performance is subject to the dynamics and liquidity of the market for the derivatives it employs to synthesize its position. These are material risks. Investors should be plainly informed of them.

    The bespoke-CDO-like structure of ABACUS is even riskier, because the derivatives required to synthesize its position are terribly illiquid and have no price transparency. That fact is definitely material to investors. The opacity and illiquidity of the (variant) credit default swaps used to synthesize ABACUS should have been clearly described, and some information should have been given about how these derivatives were priced (if only a statement that investors must understand that the instruments are difficult to price and must agree to accept a stated yield despite the uncertainty).

  10. Steve Randy Waldman writes:

    gennitydo — Thanks MUCH for pointing out my error, re which tranche ACA invested. It’s been fixed, with attribution.

  11. anon writes:

    Agree with the distinction vis a vis super senior and the lower tranches, but only to a point.

    Pricing whether “actual” or “simulated” of the unfunded equity and junior tranches will/must be implicit in the pricing of the funded tranches.

    The negotiation of pricing for the funded tranches must/will take into account the pricing of the lower tranches – that would be the case whether or not the lower tranches were funded (market priced) or not (“simulated priced”). The buyer of the funded tranches takes the applicable information set into account, and decides whether its rich or cheap relative to where he sees value in his own tranche. The buyer sees less total market pricing input on this particular deal, but has other market references, and must still make a decision on whether he “agrees” with pricing for the lower tranches – which would be the case whether or not they’re funded and market priced or not funded and simulated priced.

  12. Steve Randy Waldman writes:

    anon — Agree with you entirely on what “ought” to happen. There is an adding up constraint — the parties can access market prices to learn the total spread available from the reference portfolio, and regardless of whether tranches are funded and unfunded, they can negotiate the spread due all tranches ensuring a fair allocation in the notional CDO. Once that allocation is set, investors know what to expect for whatever tranche they fund.

    But the character of those negotiations is different for a “vanilla” structure vs a “bespoke” CDO: In a vanilla structure, the issuer is not really a party to those negotiations. The credit spread on the reference portfolio will be earned at a market rate (less the issuer’s dealer spread and other transaction costs). How spreads are then allocated is tranche warfare, subject to the constraint that all tranches must be spoken for, so some consensus must be reached. There is no “right” answer to exactly what spread is due whom. If one looks for actuarial fairness, it would depend upon how you model things.

    With a vanilla CDO, negotiations between classes of investors are fraught (however they actually occur, probably informally via an issuer trying to fill the structure), while pricing is obtained from the issuer mechanically.

    But with a bespoke CDO, the issuer effectively represents all unfunded tranches, and its economic interest is meaningfully opposed to investors in a way that differs from a vanilla structure. (The opposition of interest may be present with the vanilla structure, but it is not meaningful given the transparency of pricing.)

    So the character of negotiations is different — investors must treat the issuer as a potential adversary, unless they can rely upon a continuing relationship or reputation for fairness to customers. The issuer has a major informational advantage over typical CDO investors, so investors must negotiate with particular care and perhaps build in an extra margin of safety.

    Overall, it’s pretty important that investors understand the difference between a bespoke-ish CDO and a vanilla synthetic, and know which kind they’re investing in, if they are to be able to properly negotiate a deal. The distinction between a vanilla synthetic and a bespoke CDO is certainly material.

  13. […] is an explanation of the Abacus transaction itself. This entry was posted on 26/04/2010 (Monday) at 6:15 am and is filed under Links. You can […]

  14. M writes:

    ZeroHedge has a link from SocGen about the pricing of bespoke CDOs, not sure it will help you. I closed it quickly after opening it.

  15. frankl writes:

    i think a helpful construct to consider when advancing your understanding of abacus et al, is to consider disaster (earthquake usually, but flooding too) insurance and/or weather derivatives – there can be NO way for anyone to be long or fund the underlying, but they are legal contract bets on outcomes

  16. Jon W writes:


    To your point regarding the issuer. Isn’t ACA as the manager the one that determines the rates / structure for each of the tranches? I would assume ACA was the one that would determine what each spread over libor each tranche on the long side received, since it’s part of the strucuring.
    After all the pricing on a CDS and therefore the synthetic CDO tranches is essentially your spread over LIBOR.

    Does anyone have an idea on how these #s were determined? IF ACA determined it then how is it goldmans at fault that the price is incorrect?

  17. Mustafa writes:

    Steve your analogy of SCDOs and software implementation reminded me of this article by Joel, law of leaky abstractions. More people should read it to get a better grasp on how abstractions can fail.

  18. Jon W writes:

    Ignore my last comment. It looks like the issuer ABACUS which I assume is goldman determined the tranche spreads over libor.

  19. Sandrew writes:

    I just want to point out the distinction between the terms “bespoke” and “single-tranche” with respect to a credit derivative product. Bespoke describes the portfolio, whereas single-tranche describes the presence of attachment/detachment points (i.e. correlation risk). When the term bespoke is invoked, it means that the reference portfolio consists of a hand-picked basket of bonds. Contrast this with a deal written on the ABX.HE.BBB.2006-2 index. Single-tranche means that the credit derivative isn’t a vanilla CDS (or Pay-as-you-go CDS, in the case of structured paper) where the credit risk being traded refers to the entire loss amount (0-100) on portfolio, but rather it has explicit attachment/detachment points (e.g. 21-45).

    Apologies if this is too pedantic or if you already knew it.

  20. […] matters in the ABACUS deal.  (Interfluidity also […]

  21. anon writes:

    “tranche warfare” – love it

    “But with a bespoke CDO, the issuer effectively represents all unfunded tranches, and its economic interest is meaningfully opposed to investors in a way that differs from a vanilla structure. (The opposition of interest may be present with the vanilla structure, but it is not meaningful given the transparency of pricing.)”

    I’m not sure about that. Assuming unfunded tranches also means not issued, then the issuer has no liability, no need to hedge, and no real economic interest in them. The opposition of interest depends on the issuer’s liability (or asset) in respect of various tranches and his view on positioning/hedging, etc.

    General point:

    Why does the econoblogosphere tend to give the buy side a free ride wrt the investigation of fairness and responsibility?

  22. tom writes:

    very helpful discussion. I have two additional thoughts:
    1. from a credit risk perspective, Paulson could have pursued a similar transaction by buying protection on the comparably rated ABX index. Instead, he pursued this far more complicated structure. Since we know, from Zuckerman’s book, that he did use the ABX for some other positions, why did he not do so here? Bankers have argued that the ABX was thinly traded and therefore a $1 billion short bet would have moved prices “too much”. However, I think that is precisely the point – the true price of the risk protection would have been much more expensive if not for this structure.
    2. how was the pricing determined for the reference CDS on the MBS and on the CDOs? Since it appears that the point of this synthetic CDO was to obscure things, it seems likely to me that one of the most important things being obscured was the price of risk, not the risk itself. Goldman and other banks were coming up with structures like these that were designed to reduce price transparency, rather than increase it. Again, if Paulson shorted the comparably rated ABX for $1 bn and the other hedge funds betting against the market had done the same via the ABX, they would have overwhelmed capacity for selling protection and pushed out spreads. Obviously, this would have send a warning signal to many in the market that risk was potentially increasing. With all of these bets being disguised, instead, via deals like Abacus, how would it be possible to read what the true price was for the MBS or the CDO?

  23. Charles writes:

    What I do not understand is the following: p.22 of the prospectus, risk factors, leverage, “Through the Credit Default Swap, investors in the Notes will be effectively providing the Protection Buyer loss protection with respect to each Reference Obligation up to the Reference Obligation Notional Amount of such Reference Obligation. Losses incurred will be borne by the Noteholders. Since the Reference Portfolio Notional Amount [USD 2 billion] for the Reference Portfolio exceeds the Aggregate USD Equivalent Outstanding Amount of the Notes, [USD 192 million] investors in the Notes are providing such loss protection to the Protection Buyer on a leveraged basis.”
    Why take a USD 2 billion hedge on securities worth USD 200 million?
    Note that Paulson made 1 bio USD from his protection. So is it possible that Goldman kept the other USD 1 billion and made not only USD 15 million of fees but USD 1 billion net gain taking the same bet as Paulson??

  24. csissoko writes:

    Anon: “Why does the econoblogosphere tend to give the buy side a free ride wrt the investigation of fairness and responsibility?”

    Speaking for myself only I can tell you that I have never at any point felt that I had any input whatsoever on the prices of the financial assets I have purchased. I have no choice but to “trust” that they are fairly priced and that the markets I trade in are not manipulated. Of course, my sympathies are with the buy side.

    Whether or not “the buy side” fall into the category of sophisticated investors, they never have the same information as the investment banks/market makers and must always trust in some measure that their built in informational disadvantage is not being abused. Pretending that a market making bank and some pension fund manager who’s got a small CDO portfolio are on equal footing in a transaction is just silly.

  25. MR writes:

    Steve – There isn’t such a big difference between a single tranche synthetic CDO and a full capital structure synthetic CDO. From the perspective of the dealer i.e. GS, the risk exposure from selling a single tranche is equivalent to being the protection buyer on the individual assets’ CDS and owning all the unsold tranches (the only difference is that all the unsold tranches take up space on the dealer’s balance sheet). To take an example, if we have a synthetic CDO with CDS on 5 assets and 2 tranches only one of which is sold, it is equivalent for the dealer to
    a)either buy protection on the 5 assets from the SPV and own the unsold tranche, or
    b)just issue one tranche with a swap between GS and the SPV reflecting the cashflow payable to the sold tranche.

    Regarding pricing of the tranches, either way you’ve got an illiquid hard-to-price tranche – investors look at some broad details of the CDS portfolio like the weighted average spread, vintage etc and pricing of comparable deals and figure out what yield they need to take on any given tranche. For example, given that residual tranches are held by the dealer even in a full capital structure CDO, the investor is effectively conducting “tranche warfare” with the dealer.
    But the investor is definitely more likely to get a raw deal with a single-tranche product.

    Having said all that, the unusual thing here is that this 2-tranche CDO was hedged by GS with an identical and opposite swap on the tranches themselves with Paulson. This is not the norm. The norm is what is mentioned in the Nomura guide you have linked to – the dealer dynamically delta hedges the sold tranches and occasionally hedges higher-order risks such as correlation with other tranches, either bespoke or on an index such as the ABX. This is where Goldman’s assertion that each synthetic trade has an identical short on the other side is a little disingenuous. Most bespoke single-tranche CDOs have the dealer on the other side of course, but the dealer usually does not have an exact short as a hedge. On almost all structured products including CDOs, dynamic hedging is the norm and not static exact hedges, let alone with a counterparty who has significant input on the trade construction.

  26. csissoko writes:

    Pursuant to my previous comment, I’m not sure the “sophisticated investor” designation really makes any sense all. Why not separate market participants into price makers and price takers — possibly with multiple categories of price takers. Anybody in the “price maker” category would know that it is entirely their own responsibility to make sure that every product they buy/sell is correctly priced.

  27. Jon W writes:

    csisoko/Anon: “Why does the econoblogosphere tend to give the buy side a free ride wrt the investigation of fairness and responsibility?”

    In this situation paulson + IKB are both technically on the buyside. They’re just buying the two sides. But in reference to your question and csisokos points

    I know I’m in the minority in terms of my opinion on the relevance of leaving out Paulsons role. I personally don’t particularly have much sympathy for the buyside in this transaction. I understand that investors and consumers need to be protected from outright fraud and abuse. But there is a huge gap between someone knowing the securities are overpriced and thinking they’re overpriced. It is certainly a grey area, but there are and always will be tons of grey areas. To me the gap here is the difference between serving poisonous coffee and having to label the cup “contents may be extremely hot.”

    Sure the banks have more information, but there’s always more information to be had. Sometimes having more of it doesn’t even mean it gives you the ability to make better decisions. Many times the additional information is just noise. Even with all the information these big banks had. many of them failed. It was certainly not for lack of information. (Fact is many times the guys at these banks are making the same poor decisions you are)

    Failed banks, investors like IKB, investors in .com stocks all lose money for the same reasons.
    It isn’t lack of knowledge, but rather completely ignoring information altogether while pursuing easy returns. It’s not “I didn’t know”, it’s “I didn’t care”. Banks relying on mathematical models, ignoring risk reports, investors resting their entire decision on a Moody’s rating and people buying .com shares at insane prices are all symptomatic of the same problem. In the face of what are easy returns, careful work and responsibility take a back seat.

    You can regulate or put as many disclaimers and warnings as you want, but it won’t ever change. The same people that did it in the internet boom are the same people this time. They’ll probably do it again. It’s human nature. Everyone wants an easy handbook to getting rich, and when they’re offered one they’ll readily accept it. There’s a reason Warren Buffet is Warren Buffet, it’s because he can ignore the voice pushing him to take a short cut.

    The problem I have is this situation is pointing at one particularly questionable area of relevance. We can keep adding more disclosures, more warnings and more measures to protect people but it won’t work. All it will do is give people a false sense of security, leading them to make even more blindly informed decision.

  28. Greycap writes:


    This pass-through structure from investors to Paulson – essentially making Goldman a glorified broker – also seemed unusual to me, in the context of single-tranche synthetics. It makes sense, though, when you consider that a normal single-tranche is driven by the buy side but this one was driven by a protection buyer. Also, as I mentioned on Salmon’s blog, I don’t think ABX is a very practical instrument for hedging a correlation book. So the structure was good for Goldman too. What doesn’t make obvious sense is the 45-50 tranche Goldman held on to. I can think of three possibilities:

    1. They were originally planing to punt on 45-100 and changed their minds; then found that the price of 45-50 was more than they were willing to pay.
    2. They had originally lined up a balanced book but part of the deal fell through: variation of 1.
    3. They had a correlation hole somewhere else in the firm that the 45-50 conveniently filled.

    Any other ideas?

  29. MR writes:

    Greycap – It makes sense given Paulson was the driver in the deal. But getting ACA involved as a portfolio selection agent in such a deal doesn’t help Goldman’s case. Incidentally, Deutsche did the same deal except without a portfolio selection agent which I think helps their case.

    And on the 45-50, option 2 or 3 most likely. The fact that they lost money on one deal is irrelevant given that they are a marketmaker and they manage their book in totality. If the other mails leaked out over the weekend prove anything, it is that Goldman had no positions on their book that they did not want.

  30. Morgan_03 writes:

    Nice work Steve, i appreciate your efforts in getting it right. You, alea, macroresilience and a couple others are the only ones that took the time to lay it out correctly and not jump the gun on gut feel after the story broke.

  31. […] Desconstructing Abacus de Interfluidity […]

  32. anon writes:

    Jon W,

    I’m thinking of IKB and ACA as buy side – i.e. receiving income in exchange for selling protection.

    Agree with your sentiment.

    So now do we also throw Goldman in jail for shorting IBM or even selling IBM because Goldman has expert reseach that none of the rest of us can afford and “knows” IBM is overpriced? Where does it end?

  33. […] Deconstructing ABACUS […]

  34. Steve Randy Waldman writes:

    M — Thanks for the pointer. The article is quite interesting (although I certainly haven’t work through it in depth.) I wonder if there is any way ex post to evaluate the degree to which those models yielded actuarily fair pricing. I doubt that there is a good way, which is one of the frustrating aspects of financial theory — often not only can you be sure whether a model is predictive, but it is often very difficult to evaluate models ex post!

  35. Steve Randy Waldman writes:

    frankl — I’m not so sure. Any kind of bet can be securitized. There have existed “catastrophe linked notes”, debt instruments that pay like bonds unless there’s an earthquake or hurricane somewhere. They are issued by insurance companies, essentially as a form of securitized reinsurance. A person can be long the notes. An insurance company takes the short side, and is the protection buyer. But the insurance company is not a speculator, but is hedging its exposure to policyholders. Home-owners are naturally “long” exposure — they are well off unless a bad event happens. They take a short position by purchasing insurance to hedge that exposure. You can’t really be long the catastrophe per se, but you can be long exposure to the catastrophe, and that exposure can be hedged and transmitted with the usual tools of financial engineering. There is no one who is “naturally short” exposure to the catastrophe, that is whose risk is that it won’t happen. However, one could certainly take a speculative short position by issuing in derivatives markets or by issuing catastrophe linked notes.

    Similarly, one could certainly securitize weather derivatives. One could imagine a farmer growing a crop that dies with too much rain, and another one that dies with too little, so there would be natural exposure on both sides of the market.

  36. Steve Randy Waldman writes:

    Jon W — Your comments are always interesting, no need to ask us to ignore. Goldman Sachs had to be included in the negotiations over price, since they were going to write and be responsible for the contracts on the tranches. The question becomes who was negotiating with GS on the other side, on behalf of long investors. Did investors just accept Goldman’s marks as fair? Did ACA negotiate with GS on behalf of investors? Did IKB participate directly? Was IKB sophisticated enough to evaluate pricing of arbitrary tranches in a notional structure? Did it matter to the negotiation process or to the ultimate mark that Paulson was on the other side, presumable actively pushing for a lower spread, and that both Paulson and IKB/ACA were Goldman clients?

  37. Steve Randy Waldman writes:

    Mustafa — I love the law of leaky abstractions.

    In software, it implies we have to understand what we are abstracting over and how, and document the quirks of how we implement the abstractions.

    In finance, it implies we have to understand what we are abstracting over and how, and disclose the quirks of how we implement the abstractions.

  38. Steve Randy Waldman writes:

    Sandrew — Not too pedantic at all, and the taxonomy you describe makes intuitive sense. The sources I’ve looked at seemed to blur the distinction and use the two terms interchangeably, but it makes sense to use “bespoke” for the custom portfolio and “single tranche” for a customized pair of attachment / detachment points.

  39. […] look at what the ABACUS deal is and the potential problems with Goldman’s behaviour, see this). This action follows several derivative and private lawsuits against Goldman. It is worth pointing […]

  40. Steve Randy Waldman writes:

    anon — btw, i’m typically a short, as an investor. it’s possible to admire the Paulson / Magnetar bets while being squeamish about how they were implemented. after all, one of the main social rationales for shorting is that the practice enforce discipline and create price transparency. in theory, if there were more Paulsons earlier, the housing & credit bubble might have been milder.

    but for shorts to offer price transparency, they have to trade on markets. i can’t blame Paulson for looking for a good deal (though i’d hesitate to do business with him if he concealed his role in face-to-face meetings with ACA). but the strategy he pursued, with the ethically challenged help of Goldman and others, effectively deprived the rest of us of high commitment information that ought to have been priced into the market. if underwriters can conjure up idiots to take off-market bets, how do we expect market pricing to become accurate?

  41. Steve Randy Waldman writes:

    tom — beautifully put, a much better statement of what i was trying to express to anon above. thanks!

  42. Steve Randy Waldman writes:

    Charles — I think you are right to be confused. As far as I can tell, I think that the notional value of the reference portfolio was $2B, while its notional size if fully funded would have been $1.8B. So, in the hypothetical, fully-funded notional CDO, investors exposure would have been mildly leveraged.

    I don’t think that the $192M in actually notes sold was in any meaningful way leveraged to 2 or 1.8B. The notes of the Class A-1 and A-2 synthetic tranches would have experienced both performance and losses slightly stronger than they would have if the reference portfolio has been an “unlevered” 1.8B with respect to the notional CDO. Remember, what the notes experience has nothing to do with the quantity sold. Its performance is based on what a hypothetical fully funded tranche would have experienced in the notional CDO. If the Class A-2 tranche would have taken a 50% hit in the CDO, so will ACA and IKB’s notes, however much they had invested.

    Throughout the prospectus, the distinction between when things are real and when they are notional is blurred. The ABACUS is the least clear investment document I have ever encountered. It does not meaningfully explain how the structure behaves unless you bring a lot of inference and assumption to it.

  43. Steve Randy Waldman writes:

    MR — Great point on the equivalence between a single tranche synthetic and an ordinary deal where the dealer takes some of the tranches. As you say, in that situation you are again in potentually adversarial negotiation with the better informed dealer.

    I think in practice there is still a difference though, because usually dealers took the super senior tranche and just that. I’d imagine that pricing of the more junior — equity, mezzanine — tranches would be more uncertain and contentious. Also, the degrees of freedom through which the dealer can exploit his informational advantage increases with the number of tranches he is (literally or virtually) taking. If only a single tranche is being sold to the end client, but spread is being allocated to several notional tranches (perhaps just one tranche more junior and one more senior than the synthesized tranche), the dealer can take an extra sliver in multiple places that the buyer may be unable to detect.

    Ultimately we’re agreeing, I think. As you say, investors are a bit more likely to get a raw deal.

    In this case, where, unusually, there was an active player on the other side negotiating against the two funded tranches of the deal, there might be even more hazard.

  44. Steve Randy Waldman writes:

    Jon W / csisoko — There’s ordinarily something of a balance between the “buy side” — meaning liquidity demander — and the “sell side” meaning market makers or liquidity providers. On the one hand, the market makers are professionals, intimately involved in the markets, the builders and modelers of these complex structures. They are price makers.

    But, the players in the “sell side” must compete to provide liquidity on demand. That is, they are expected to take the other trades that “buy side” clients initiate. So even though the average market maker might be better informed and more sophisticated than the average buyer or seller of an instrument, the “sell side” players must always transact with clients who may be better informed than them about a particular security or position. The ability of the “buy side” to select the timing of trades with no duty to disclosure their information makes the game more even. Most “buy side” players, of course, have no special information, and the liquidity of markets is largely a function of the ratio between informed and “noise” traders. When there are lots of noise traders, the “sell side” is less nervous and provides inexpensive liquidity. When there are informed traders, the “sell side” gets defensive, sets spreads wide, and adjusts prices quickly.

    What bugs me so much about this deal is that hinted at in Jon W’s comment that both Paulson and IKB were on the “buy side”. That sound wrong, no? Paulson actually initiated the trade, and IKB provided liquidity. IKB unwittingly took on the role of the “sell side” liquidity provider, without the information (crucial to the real “sell side”) of who had initiated the trade.

    Ultimately, this was Goldman’s sin. Goldman, on the “sell side”, was asked for liquidity by Paulson. It was unwilling to provide that liquidity at its own risk, which is always its prerogative — market makers can decline to offer liquidity, albeit at a reputational cost. But rather than bear the hit to organizational pride, it used its role in the industry as a trusted unerwriter to conjure a party to supply liquidity and take on the risk in its place, without information that Goldman itself would require as a market maker (who initiated the trade).

    I know I’ve mangled the actual story line a bit, in that Paulson suggested the scheme, rather than asking for liquidity on a scale he knew Goldman would not provide at the price he wanted. But it was Goldman that had to implement it, that conjured and sold a deal to customers who were taking on market-maker inventory by a potentially informed trader without knowing that or understanding what that meant. Even a “sophisticated” buy side investor doesn’t know the art of market-making, especially when it doesn’t even know it’s in the role.

    Jon W — In a comment to a previous post, you asked how investors would like it if sellers changed their prices against them as soon as they tried to buy. But of course they do! A market maker’s spread is an automatic price adjustment from the midpoint price, depending on whether the initiator of the trade wishes to buy or sell. A big spread means the market maker is nervous that the guy initiating the trade might know something. (it could also mean a thin issue, but holding general trading frequency constant.) The depth of the order book (or lack thereof) defines the terms under which liquidity providers can revise pricing based on order flow. Liquidity provision is a facsinating game, and its largely about watching who is initiating and tracking flow. Liquidity providers take these factors as markers of risk, and adjust their prices and willingness to take inventory accordingly. It’s what Goldman is in the business of doing, but IKB was sweet talked into providing liquidity blind.

  45. […] Filed under: Brief Comments | Tags: CDO, CDS, Toxic assets | Steve Waldman has a great post  up deconstructing the Abacus CDO that was the source of the SECs charges against Goldman.  What has become clear is that Abacus was […]

  46. […] suspected this, but I haven’t had the time to look at the marketing documents. But thankfully Steve Randy Waldman did. I don’t think I can improve on his description — these things take hundreds of […]

  47. anon2 writes:

    There’s way too much attention being given to the structure of this deal. Review all the details and the structure is not only sound but well understood by institutional investors. If an investor didn’t understand the structure they shouldn’t have been involved. There are thousands and thousands of these deals in the market, there are others that are vastly more complicated. With the appropriate background knowledge most of these deals can be explained informally in a page or less.

    What is lost in all this is some perspective on the way the world worked when investors were contemplating this purchase. Essentially if a security carried a AAA it was subject to very small capital requirements and therefore financial organizations were economically encouraged compensated to purchase AAA rated securities. The fact that someone would go to all the trouble to create a hypothetical security is an indicator of the insatiable demand that existed for anything with yield at the time.

    This CDO actually has, in the case of some of the underlying holdings, up to 3 layers of leverage. The leverage in the CDO was not accounted for correctly along with the correlation between the mortgages ultimately issued in the underlying issues. Ratings agencies used the ratings they assigned to the underlying securities (without significant regard to what the securities were other than that they were diversified) to create ratings for the CDO.

    Want someone to be upset with for the severity of the crisis? Rating agencies and rule based regulation are far more culpable than market participants. Want to be upset about the root cause housing crisis? Turn your ire towards mortgage brokers and low doc/no doc loans.

  48. Moe writes:

    I’m totally new to this, but my sense is that what we actually think of as a synthetic CDO was usually termed a “hybrid” CDO because it comprised (say) 80% synthetic MBS and 20% cash bonds?

  49. […] The 196 page prospectus. Deconstructing Abacus and more links about it. […]

  50. Great analysis of the structure of the Synthetic CDO here. Defnitely enjoyed the link to the CDO waterfall visualizations and will probably incorporate in our site.

    What concerns me is that everyone has found this chink in the armor of GS and therefore it’s time to crucify them. To me it seems a lot like a populist witch hunt demanding that the rich and profitable accused be released to the masses, so that they might have their vengeance. It wouldn’t be the first time that one was sacrificed unjustly to calm the emptiness among a collective soul.

    The author claims that basically these “sophisticated investors” weren’t “sophisticated” enough to sit down at a table and negotiate a price for the ABACUS tranche they were interested in. However, all the buyers needed to know was that they were buying a security that was fully linked to no real assets, and that this sort of activity was surely risky. Buyer beware. Or more aptly, greedy buyer get what you deserve when you buy something that’s too good to be true. These things might have beeen labeled AAA, but they weren’t rated by an agency and weren’t tied to actual cash flow producing assets. Besides, plenty of the vanilla AAA CDO tranches were torn up as well.

  51. […] on the Goldman ABACUS deal have been written on the interfluidity blog; I recommend the posts of April 25,  April 27,  April […]

  52. […] tip to Baseline Scenario, Alea Blog, and Interfluidity for their analysis of ABACUS, which helped us bring you some of the links and data that contribute […]

  53. […] In a series of posts on Goldman’s Abacus CDO, Steve Waldman discusses another case of welfare reducing structured […]

  54. […] Kwak was himself basing his argument on this post from Steve Randy Waldman, who hasn’t got his head around this ABACUS thing. How can be there […]

  55. […] In the various blogosphere efforts to dissect the Goldman Abacus transaction now in the SEC’s crosshairs, some commentators have characterized it as unusual, a “bespoke” CDO“, or “a very complicated deal…a supersynthetic CDO.” […]

  56. JKH writes:


    I agree with your May 3 update comment re the Naked Capitalism comment.

    The emailer notes the unfunded nature of the super senior tranche, but misses your more general and fundamental point regarding the unfunded, absent nature of a number of other well defined tranches.

    As you’ve described, this prohibits the normal comprehensive waterfall construction for cash flows fully replicating what would be determined by the reference asset portfolio. It effectively forces the construction of select waterfall subsets – which necessarily means directly synthesizing the cash flow of chosen liability tranches – cash flows subsets determined according to the reference portfolio and the reference waterfall structure.

    Difficult stuff to describe. Not sure I’ve succeeded.

  57. Young Gun writes:

    As a young, 21 year old investor, I find myself drawn to drawn to this news story as an opportunity to learn. Seeing how the discussion on this blog is the most intelligent one I have discovered to date, I would like to present my thoughts to be critiqued by the community. Please note that I am disregarding morality in my analysis; I am only concerned with the legal facts of the case.

    My overall position is that the SEC realized that it was (literally) jerking off during the financial crisis, knew that the Obama administration would be forced to react by populist sentiment, decided that Goldman was the easiest political target, and targeted a highly complex deal completed at the very end of the crisis. The defenses I would present in Goldman’s favor have been oft repeated so I won’t rehash them here. That being said, I do have some strong concerns about the case.

    My concern is about the apparent assumption held by ACA that Paulson was intending to invest in the equity (or FL) portion of the synthetic. My question is multi-faceted; is the described “equity ” portion of the deal equivalent to the “first loss” tranche described in the circular offering (I’ve noticed that Goldman reps always describe Paulson’s interest in the FL tranche while the SEC talks about an “equity” portion), if the circular offering shows that the more Junior tranches will remain unfunded does it matter what ACA thought Paulson’s interest was, and could material damages be proven in the wake of a broad, market collapse?

    My personal thoughts are that Goldman allowed ACA to assume that Paulson was going long and skirted the issue when question. ACA failed in its responsibility to do proper due diligence by not pressing when Goldman was reluctant to pin down Paulson’s exact interest. Furthermore, there is a section in the circular offering where ACA describes its own role as the portfolio selector. Nowhere in that section does ACA describe the role of Paulson in the selection process. Also, both parties (ACA and IKB) had a desire to be long MBS type risk and would have experienced massive losses no matter what investment vehicle they chose. Finally, there are boilerplate disclosures in the documents the explicitly state that Goldman may have knowledge of material, non-public information that has no duty to disclose.

    The ultimate question is, why was there confusion over whether Paulson would hold an “equity” tranche if it was never ultimately funded and does that matter if the reference portfolio is rubber stamped by ACA? Was it an outright lie? Did Goldman allow ACA to assume? Or did Paulson actually have some exposure to that risk that he immediately offloaded. I suspect that option number 2 is true. No matter what, I think that Goldman is on track for a settlement as no prosecutor would want to hang his reputation on a jury to decision in a case like this. In fact, I think Goldman comes out of this situation stronger, with a more client-oriented attitude as these investigations will put the fear of God into upper management.

    Please comment on my analysis. I always appreciate feedback.

  58. Justin Case writes:

    Are all Hybrids dangerous? If so, should I sell my Toyota Prius?

  59. […] particularly applies to so-called synthetic securities like the Abacus structure. Pricing these securities, part of that constellation described by Mr. Buffett as “weapons of […]