Everyone knows the story of Abacus 2007-AC1 by now: Goldman Sachs sold some mortgage-backed-security CDOs to some people, and those people thought that the underlying mortgage-backed securities were chosen by an outfit called ACA Management to be Good, but in fact they were chosen by Paulson & Co. to be Bad, and they turned out to be Bad, and that was Bad. The SEC sued Goldman over it, and Goldman settled for $550 million, and then everyone else sued too because they had been lied to about who picked the mortgage-backed securities (Paulson, not ACA) and why (to fail, not to succeed).
Among the people who sued was ACA, whose role in the transaction was (1) pretending to pick the underlying RMBS and (2) issuing a financial guaranty policy (to Goldman) referencing the super senior tranche of Abacus. That tranche more or less went poof, and ACA ended up owing $840 million to Goldman (though, really, ABN Amro paid the $840mm, and Paulson got it).1 Since ACA was in the business of writing terrible financial guaranty policies, it blew right up and ended up paying only $30 million. Then it sued Goldman for the $30 million back, plus punitive damages. ACA's claim is that, while it knew that Paulson had selected the underlying RMBS, it thought Paulson was net long Abacus, because Goldman schemed and lied, and that it wouldn't have insured Abacus if it'd known the truth about Paulson's position.
Yesterday ACA lost when a New York appellate court dismissed its case. The court split 3-2, and the opinion is short and pretty weird; basically the majority says "it doesn't matter that Goldman lied to ACA about Paulson's position, because ACA should have kept asking until it got the truth," which is a funny law.2 The two dissenting judges seem to have rather the better of it.3
Still the result seems right. I hope from my little summary it is clear that ACA is perhaps the least sympathetic of all the financial crisis victims. From what they claim themselves, they allowed Goldman to lie about their role in Abacus, signed off on a misleading offering circular, paid off their debts at 3.5 cents on the dollar, and then sued for the 3.5 cents back (plus punitive damages).
Like: the gist of what Goldman did wrong was that they told everyone that ACA had picked the underlying MBS (to succeed!) when in fact Paulson had (to fail!), but keen observers at ACA might have noticed that this wasn't true because they hadn't picked the underlying MBS. Here's how ACA describes its role in their complaint:
ACA Management, LLC, a wholly owned subsidiary of ACA ("ACAM"), was the pro forma portfolio selection agent for ABACUS, i.e., ACAM agreed to and relied upon a portfolio largely selected by Paulson.
Here's how the Abacus offering circular, which ACA presumably reviewed and signed off on, describes ACA's role:
The Initial Reference Portfolio will be selected by ACA Management, L.L.C. ("ACA Management" and in such capacity, the "Portfolio Selection Agent") pursuant to the terms of the Portfolio Selection Agreement, dated as of the Closing Date (the "Portfolio Selection Agreement"), between the Issuer and the Portfolio Selection Agent.
The words "pro forma" do not appear in the offering circular. (Nor does the word "Paulson.") On the other hand, the Abacus pitchbook does include a whole section on ACA's portfolio selection role, including a page on its investment philosophy that says things like "Asset selection and asset management premised on credit fundamentals and then optimized for relative value" and "ACA Management will utilize proprietary models to stress and confirm the adequacy of cash flows." There's no bullet point for "Meh, they just trust Paulson." I have some sympathy for the investors who bought "designed-to-fail" CDOs without doing their own credit work, but that sympathy breaks down when the investor is the portfolio selection agent who was hired to do the credit work.
You can if you like criticize the New York court's ruling for its rather strict buyer-beware standards, which would prevent someone who was lied to from suing unless they kept asking about the lies a lot. But whereas it might not be appropriate for everyone, or even for every "sophisticated counterparty," it's surely appropriate for ACA, who were as much co-conspirators as they were victims. They knew with perfect certainty and clarity that Goldman was misleading investors about how the Abacus portfolio was being selected, because they were helping Goldman to do the misleading. They knew that every representation Goldman made to everyone else about the portfolio selection was at least a little off. That makes the fact that they nonetheless assumed that Goldman would never lie to them harder to justify.
ACA Financial Guaranty Corp. v. Goldman, Sachs & Co. [NY Courts]
Judge Says Insurer Should Have Known Better on Debt Deal [WSJ]
In tossing ACA case vs Goldman, N.Y. appeals court says buyer beware [Reuters]
1.That's from ACA's complaint, paragraphs 111-113. Note (paragraph 89) that ABN Amro actually intermediated the ACA financial guarantee, basically because Goldman thought ACA was a crap credit: "In effect, ABN assumed the risk that ACA would default on its financial guaranty policy, while Goldman Sachs insulated itself from the risk that ACA would default on its financial guaranty policy." ABN actually paid out the $840mm (and got ACA's $30mm). Similarly Goldman was in effect intermediating to Paulson, since Paulson bought protection on the super senior tranche from GS that more or less back-to-backed the protection that GS bought from ACA (through ABN Amro). So the money went:
ACA -> ABN Amro -> Goldman -> Paulson
Except ACA didn't pay its share, so ABN Amro was the ultimate loser (and Paulson the ultimate winner).
2.What they actually say is:
In sum, plaintiff's fraud claims based on the allegation that plaintiff, a highly sophisticated commercial entity, was misled into believing that a nonparty hedge fund would take a long position in the first-loss tranche of the collateral debt obligation, in alignment with plaintiff's interests, must be dismissed because: (1) such misrepresentations were specifically contradicted by the offering circular's disclosure that no such equity position was being taken; (2) plaintiff's alleged reliance on such misrepresentations would have been contrary to its acknowledgment (as set forth in the offering circular) that, in entering into the transaction, it was "not relying (for purposes of making any investment decision or otherwise) upon any advice, counsel or representations (whether written or oral) of [defendant] . . . other than in the final offering circular for [the transaction] and any representations expressly set forth in a written agreement with such party," and that defendant was not "acting as a fiduciary or financial or investment adviser for the purchaser"; and (3) the hedge fund's intentions with regard to this investment were not peculiarly within defendant's knowledge and plaintiff, although it was in direct contact with the hedge fund, failed to ask the hedge fund what position it intended to take in this investment.
Point (1) we'll discuss a bit further down. Point (3) is just "they could have kept asking after they'd been lied to a bunch." Point (2) is kind of an important one; a lot of assets are sold by means of:
- Seller calls Buyer and is like "this asset grows magic beanstalks, you should totally buy it," or whatever;
- Buyer conducts due diligence in which Seller representatives talk a lot about magic beanstalks;
- Buyer signs contract to buy the asset;
- the contract represents "this asset is some beans" but is silent about magic beanstalks; and
- the contract contains an acknowledgment (an "integration clause") that Buyer is not relying on any representations other than those specifically included in the written contract.
You sort of can't have M&A, securities sales, etc., without being able to rely on that acknowledgment in almost all circumstances, so it's perhaps a good thing that the court gave it a lot of weight. Still sometimes fraud outside the contract might matter, no?
3.Though here's a fun point in the majority's opinion:
[P]laintiff [ACA] received, inter alia, the offering circular for the transaction, which expressly disclosed that no one was investing in the first-loss tranche. This information should have alerted plaintiff that contrary to the representations made, the nonparty hedge fund [Paulson] was not funding a portion of the transaction at all, let alone in the manner represented (i.e., by taking the equity or long position). Therefore, plaintiff should have questioned defendant [Goldman] or the non-party hedge fund; such an inquiry would have likely informed plaintiff that the nonparty hedge fund was taking a short rather than a long equity position represented.
Ha it's true, the Abacus offering circular specifically discloses that zero dollars worth of Class FL Notes (the 0-10% first-loss/equity piece) were sold. That does look like it should have alerted ACA to the fact that, y'know, Paulson didn't buy any, even though Goldman had told ACA that he would. Nonetheless the dissent argues that he could have been long the equity synthetically:
The complaint alleges that long investors can participate in the capital structure of a synthetic CDO such as ABACUS either by purchasing notes or by selling protection on a specified tranche in the capital structure. Given this description, there is a reasonable inference that plaintiff understood the absence of equity notes to mean that Paulson intended to "take a long position in the equity tranche of ABACUS through a [credit default swap]," by selling protection on the 0-10% tranche instead of purchasing notes.
I'm not sure how likely that is, though the majority's response is essentially gibberish:
We reject the dissent's assertion, that the absence of any funding of the first loss tranche was attributable to the fact that the non-party hedge fund was purportedly funding the first-loss tranche by taking the long position on a credit default swap. This assertion does not explain why the tranche was completely unfunded, since even the funding mechanism perceived by plaintiff - the credit default swap - should have had a value and thus should have been listed in the offering circular.
What? No. That's not how it works.