Deconstructing Goldman Sachs's Fraud Defense
In short, Goldman argues that:
- Paulson & Co. didn't get to stack the deck, because both major purchasers, IKB (a large German bank) and ACA Capital Management (which figures heavily in the SEC complaint) had input into the choice of securities, with ACA having ultimate control.
- Goldman didn't deceive ACA about Paulson's interest in the deal.
- Goldman fully disclosed the mortgage securities involved in the Abacus synthetic collateralized debt obligation (CDO), and the sophisticated investors involved all knew synthetic CDOs have a long and short side.
- Goldman took a long position in Abacus, which caused it $90 million in losses, dwarfing the $15 million it made putting the deal together.
What Was Paulson & Co.'s Role in Selecting the Abacus Securities?
First, it's a key factual dispute as to how much input into securities selection for the CDO that Paulson & Co. had. If Paulson didn't have undue influence over the securities selected, the SEC has no case because there would be nothing important to disclose about Paulson's role, no material omission -- and thus no fraud.
The SEC's complaint doesn't mention any input from IKB, which Goldman asserts. So, let's set aside IKB because there's really no way to assess this one at the moment, and focus on ACA.
Goldman notes that ACA was a major buyer of the CDO and thus evaluated the securities with an eye toward their appropriateness as long-term investments. Goldman also says that as "Portfolio Selection Agent," ACA had final authority to determine what securities the CDO referenced. Sounds very good for Goldman so far: Paulson was negotiating with its counterparty, and its counterparty had final control. In that scenario, it doesn't seem like Paulson has the juice to rig the deal.
So what does the SEC say?
The commission's complaint, at paragraphs 25 to 35, tells its version of Paulson's role in selecting the portfolio. The SEC focuses on communications among Goldman, ACA and Paulson, noting that Paulson hosted the initial meeting among the three companies on Jan. 8, 2007. The next day, Goldman sent ACA an email with the subject line "Paulson Portfolio" with an attached list of 123 securities, presumably all chosen by Paulson. On Jan. 10, Goldman Vice President Fabrice Tourre wrote to ACA re: "Transaction Summary." "[W]e wanted to summarize ACA's proposed role as 'Portfolio Selection Agent' for the transaction that would be sponsored by Paulson (the 'Transaction Sponsor'). . . . [s]tarting portfolio would be ideally what the Transaction Sponsor shared, but there is flexibility around the names." So far, it seems that the selected portfolio began as Paulson's deck, with Goldman's full support.
Nearly two weeks later, ACA replied with its own list of 86 securities, noting that 55 (64%) of them were from Paulson's original list. Negotiations on the portfolio continued, with ACA, Paulson and Tourre meeting on Feb. 2. During the meeting, Tourre wrote an email describing the meeting as "surreal," but he didn't explain why. The SEC suggests it was surreal because Tourre and Paulson knew that Paulson wanted to include securities that will lose value, and ACA didn't know that.
After the meeting, ACA circulated a list of 82 securities that ACA and Paulson had agreed on, plus 21 "replacement" securities, and asked for Paulson's approval of the list. Three days later, Paulson rejected eight but left the rest, which somehow resulted in a list of 92 securities. ACA then discussed internally whether or not to sign off on the list, noting that the final CDO would reference anywhere from 80 to all 92 securities. Although ACA agreed to the list, it wondered why Paulson rejected Wells Fargo securities because Wells Fargo subprime mortgages were perceived as particularly sound. Two days later, ACA and Paulson settled on a list of 90.
If this version of events is accurate -- and on the one hand, it's backed by documents but on the other, the documents cited by the SEC are out of context and might read differently in context -- the portfolio-selection process seems to be a joint effort by Paulson and ACA, one in which ACA believed it needed Paulson's approval of the securities and which started with Paulson's initial wish list. Goldman, which hired ACA, appears to have urged ACA to accept Paulson's securities as much as possible.
Nonetheless, Paulson's influence wasn't total; ACA didn't take dictation from Paulson. Moreover, ACA obviously scrutinized the securities being included, and if the portfolio was clearly designed to fail, presumably ACA would have noticed. ACA was a sophisticated player that had done over 20 prior deals. The fact that Paulson wasn't drinking the then-current mortgage-security Kool-Aid and thus did a better job of analyzing the securities doesn't mean that at the time anyone other than Paulson, or those handful of investors who thought like it, would look at the CDO and think the deal was rigged.
Indeed, it's possible that if potential buyers of the CDO had been explicitly told Paulson had tried to rig the security, perhaps they would have scoffed at the claim, asserting that Paulson was crazy and would happily bet against him. (On that note, Paulson has said, "the vast majority of the people in the market thought we were dead wrong and openly and aggressively purchased the securities we were selling.")
Of course, if everyone had been told of Paulson's role and the worldview motivating it, the SEC wouldn't have filed its suit. Because ultimately it does not matter if Paulson had enough influence to truly rig the CDO. It just had to have enough influence that investors would have found it really important to know, before buying, about Paulson's role. If that's the case, then failing to tell investors about Paulson's role and interests was a material omission, which is a necessary element of fraud (see more below on that). Advantage so far: the SEC.
Did Goldman Deceive ACA About Paulson's Interest?
For two reasons, the SEC's case in large part hinges on showing that Goldman deceived ACA. First, that would show consciousness by Goldman that Paulson's anti-CDO investor interest was very important to ACA and thus to investors (ACA was also a major investor). That would support the claim that not disclosing Paulson's interest -- and in the selling documents, Paulson's role in selecting securities -- was a material omission, one of the elements the SEC has to prove to win. Second, if Goldman deceived ACA about Paulson's interest, it also supports the idea that Goldman had the intent to deceive, which the SEC also has to prove to win. Goldman flatly denies deceiving ACA about Paulson's interest. So what does the SEC say?
The SEC lays out its argument at paragraphs 44-51. The commission cites an ACA email to Goldman on Jan. 8, 2007, following its initial meeting with Paulson, in which ACA says ". . .it didn't help that we didn't know exactly how they [Paulson] want to participate in the space. Can you get us some feedback?" The SEC then cites a different part of Tourre's Jan. 10 "Transaction Summary" email mentioned above, which referenced a ""Contemplated Capital Structure" with a "% - %: pre-committed first loss" as part of the Paulson deal structure." The SEC asserts that this description of a "% - % tranche at the bottom of the capital structure was consistent with the description of an equity tranche." If Paulson were investing in the equity slice of the CDO, then it was on the same side of the transaction as ACA, and their interests would have been aligned.
Two days later, Tourre had a phone conversation with ACA about the deal, and two days after that, ACA emailed Goldman re:"Call with Fabrice [Tourre] on Friday." The email said in part: "...the structure looks difficult from a debt investor perspective. I can understand Paulson's equity perspective but for us to put our name on something, we have to be sure it enhances our reputation." That email was forwarded to Tourre on Jan. 16, meaning by that point, he was notified that ACA saw Paulson as an equity investor. Perhaps that's why he thought the Feb. 2 meeting he sat in on between ACA and Paulson was "surreal."
The SEC cites internal ACA documents from Feb. 12 to show that ACA continued to misunderstand Paulson's position as the deal continued. On that date, ACA approved its own role as portfolio selection agent. The ACA memo summarizing its approval of becoming the selection agent and related meeting notes describe Paulson as an "equity investor" twice. The memo says Paulson as "the hedge fund equity investor wanted to invest in the 09% tranche of a static mezzanine ABS CDO backed 100% by subprime residential mortgage securities," a description that matches the information in Tourre's Jan. 10 email. The meeting notes show discussion of "portfolio selection work with the equity investor."
On this record, it looks like the SEC again has the advantage. ACA had the idea that Paulson was an equity investor on terms identical to ones laid out by Tourre. These terms were laid out in response to a request for clarification about Paulson's interest, and Tourre and at least the original recipient of ACA's email were aware of the misimpression early on. Moreover, ACA seems concerned about what Paulson's role is, and about reputational risk. It wants to be sure the deal is OK from a "debt investor perspective."
Perhaps by putting the documents in context, Goldman will be able to show that Tourre did not intend to mislead ACA and didn't know ACA was misinformed. One potent fact on Goldman's side is that the deal apparently didn't contain an equity slice, and that information was in the offering materials. Does Goldman have a duty to disabuse ACA of a misimpression when the corrective information is readily available to it?
Of course, if Goldman deceived ACA, it also goes to whether or not Goldman intended to deceive the ultimate purchasers of the CDO as well, both because ACA assumed almost $1 billion in risk associated with the CDO's performance -- risk guaranteed by Dutch bank ABN Ambro -- and because it would show consciousness that disclosure of Paulson's role to investors was problematic and needed to be avoided. Since intent to deceive is a necessary element of fraud, the did-Goldman-deceive-ACA issue is crucial.
Is Goldman's Silence on Paulson a Material Omission?
The standard for a material omission is: "a substantial likelihood that the disclosure of the omitted [information] would have been viewed by the reasonable investor as having significantly altered the 'total mix' of information made available." The SEC alleges that not telling investors that Paulson had a significant role in choosing the securities and that it had economic interests opposite to investors was a material omission. The SEC highlights the fact that the offering materials stated that ACA selected the securities and that ACA had economic interests aligned with those of investors.
Goldman's defense asserts that it fully disclosed the mortgage securities referenced in the synthetic CDO; that sophisticated investors -- and with Abacus being a private placement, the investors were by definition sophisticated -- know a synthetic CDO has a long and a short side; and that normal business practices don't include disclosing who the counterparties are.
To spell out Goldman's case on this point more thoroughly than the firm's Friday statement goes, assume that Goldman provided all the relevant and important information about the securities referenced in the CDO it sold for Paulson & Co. such that sophisticated investors, or experts hired by them, could accurately assess the risks associated with the underlying securities.
If that's the case, then why does it matter that Goldman didn't disclose Paulson's role or positioning? That is, if Paulson could look at the securities and determine that they were likely to fail, as the SEC complaint alleges it could and did, and Goldman included the same data that swayed Paulson's opinion in the offering materials so that investors had a real opportunity to reach the same conclusion, then purchasers of the CDO could accurately assess the CDO's value before purchasing it.
Wall Streeters are supposed to make money by doing better analysis of the risks and benefits of transactions than their counterparties. All that's required is access to the same information. Given that, plus the awareness that there's always someone in Paulson's position, on the other side of the CDO, it's hard to see how knowledge of Paulson's involvement and interests would "significantly alter the 'total mix' of information made available."
The law doesn't coddle investors, just as it doesn't coddle people who took out mortgages without fully understanding the terms or carefully reading the documents involved. Just as it's no defense to a foreclosure for a homebuyer to say, "Hey, I didn't understand the mortgage I signed up for, I didn't read it, I relied on the mortgage broker's translation," it should be no defense to say, "Hey, I didn't understand what the CDO was that I bought, I relied on the ratings agency's assessment, I relied on ACA's reputation and role."
Even if the CDO was too complex for all but true experts to analyze accurately, nothing stopped investors from hiring such experts to do due diligence. To find Goldman's secrecy about Paulson's role to be a material omission would relieve investors of their buyer-beware duty to perform their own evaluation of the security they're buying.
That's a potent argument. I'm not convinced by it, but those who have a very formalistic view of how the world does or should work -- perhaps including some judges -- may well be convinced.
I reject the argument, first, because it presumes a world that does not exist, a world that Goldman itself, if the allegations are true, counted on not existing. That is, Goldman allegedly involved ACA as an independent portfolio selection agent precisely because it believed investors would rely on and be reassured by ACA's involvement, making the CDO marketable. Without such a selection agent, Goldman didn't think it could get the deal done. (SEC complaint at paragraphs 23-24.) Goldman also knew that not every portfolio agent would "agree to the type of names [of securities] Paulson want[s] to use" and put its "name at risk. . .on a weak quality portfolio." (SEC complaint at paragraph 21.)
Goldman apparently considered ACA's involvement so important that it allegedly misled ACA about Paulson's investment position to keep ACA a part of the deal, as discussed above. Goldman's offering materials also stressed that ACA picked the portfolio and that ACA had economic interests aligned with those of investors. (SEC complaint at paragraph 38.) If Paulson's role wasn't material to investors, why not just say that what Paulson's role in selecting the securities was, and that its interests opposed those of investors?
Moreover, as the investigations of the ratings agencies reveal, banks such as Goldman "ratings shopped" and otherwise attempted to get ratings agencies to give their securities the highest possible ratings, again because the banks understood that such ratings were important to investors, who at least to some degree relied on by them.
In short, Goldman's accurate perception of how the world works -- investors, including sophisticated investors, rely at least in part on proxies' due diligence and on the proxies' reputations and roles -- was crucial to its plans to sell the CDO. Indeed, it was also crucial to Paulson's strategy. A Paulson employee explained the reason Paulson was able to make the bets it did: "In my opinion this situation is due to the fact that rating agencies, CDO managers and underwriters have all the incentives to keep the game going, while 'real money' investors have neither the analytical tools nor the institutional framework to take action before the losses that one could anticipate based [on] the 'news' available everywhere are actually realized." (SEC complaint at paragraph 17.)
If, as alleged, Goldman knew the CDO was being designed to fail, and the world worked the way Goldman's-we-disclosed-the-securities-defense assumes it does (meaning that purchasers wouldn't rely on ACA or the ratings agencies and would accurately evaluate the securities underlying the CDO), then Goldman never could have done the deal. No one would have bought the CDO because everyone would have figured out how risky it was.
The second reason I reject the argument above is the incentives created by the different results. If failing to disclose Paulson's involvement -- failing to disclose that the CDO was allegedly being set up to fail as much as possible by the party who benefited if it did fail -- is not a material omission for the reasons laid out above, then we've blessed and thus encouraged the practice. I think, and I'm willing to bet most Americans agree, that it's just morally wrong if you sell people something designed to fail, without telling buyers that it has been designed to fail, so that someone who hired you to sell it can make a lot of money. It's a classic kind of snake-oil charlatanism. On the other hand, if failing to disclose Paulson's involvement is a material omission, then the practice will stop.
Note, the material omission isn't that Paulson, or anyone else, is betting against the CDO, or that even that anyone assisted ACA, or even partnered with ACA, in selecting the securities. It's that someone who is betting against the CDO assisted/partnered with ACA in choosing the portfolio. It's a pretty narrow circumstance, and it won't relieve investors from doing their due diligence in traditionally structured transactions. While the situation in this case surely wasn't unique -- apparently at least Deutche Bank did similar deals with Paulson, although other banks turned the firm down -- it wasn't the norm. (It may not have been unusual either, however, given that a 2008 article about synthetic CDOs casually says, "Many investors had deals structured to order, with their own choice of portfolio names and tranche level.")
What About Goldman's $90 Million Loss Claim?
Goldman's appeal to a common-sense defense is, "Hey, if we had deliberately structured the deal to fail, we wouldn't have held $90 million of it and thus lost $75 million more than we made by doing the deal in the first place." However, according to a December 2009 New York Times piece on the family of synthetic CDOs that Abacus belongs to, Goldman made lots of money by betting against those CDOs. Perhaps Goldman's assertion of the $90 million loss is as narrowly framed as the SEC's claim that Goldman made $15 million in fees on the deal. I have no idea if it's true, but nor do I have confidence it's not.
Beyond the possibility that Goldman actually netted a profit on the deal if the frame is set widely enough, it's also possible that Goldman had sufficient incentives to do the deal as described by the SEC while still risking a big loss by retaining a substantial interest in it. I confess I don't understand enough about the synthetic CDO market, or, frankly, the overall dynamics of the various mortgage-backed securities permutations to be able to follow the money well enough to know for sure.
However, it seems that banks routinely did CDOs, and other deals, knowing that they would retain significant risk (as Goldman did in this case), or they actively purchased risk-filled CDOs. That's because doing CDOs was incredibly lucrative for the banks and their employees alike. That is, they were lucrative until the market crashed, and all the banks needed bailing out -- in part because of their CDOs.
Apparently the case was controversial within the SEC; the commissioners approved the enforcement action by a 3-2 vote along party lines, with Mary Shapiro, an Independent appointed by President Obama joining the two Democrats in voting to bring the case. In part because the vote was party-line.
It's not clear what it means about the SEC's sense of its case. Maybe it means that the SEC isn't sure it's a strong case; maybe it means that the division is ideological. Perhaps the Republicans believe the we disclosed-everything-about-the mortgages-so-there-can't-be-a-material-omission argument; perhaps the Republicans are worried about the political ramifications for financial reform.