The latest shoe to drop on the Goldman front is the report on Wednesday that the SEC was investigating yet another one of its synthetic CDOs, this one a $2 billion confection called Hudson. It isn’t clear whether the SEC will file charges, but this one has the potential to be particularly damaging in the court of public opinion, since this CDO was created solely as a proprietary trading position to help the firm get short subprime risk in late 2006, when the market was clearly on its last legs.
By way of background, the assets in a synthetic CDOs are credit default swaps. In the case of Hudson, they referenced $800 million of BBB subprime bonds, 2005 and 2006 vintage, and $1.2 billion of the ABX. The deal was a wipeout.
What makes Hudson different from the Abacus CDO that is the subject of an SEC lawsuit is that it was not even arguably intermediated between customers. Goldman was not only the initial short counterparty (as was indicated in the contract as standard verbiage), it was every and always intended to be the ultimate short counterparty. Why does this matter?
Synthetic CDOs were sold to investors as the economic equivalent of cash CDOs, ones whose assets were subprime bonds rather than credit default swaps. That was always more than a bit disingenuous. Cash CDOs had for some time been the way that underwriters would dispose of the pieces of subprime bonds they were unable to sell, namely the riskier tranches. Conceptually, it was like taking unwanted parts from (presumably) healthy pigs, grinding it up with a little bit of better meat plus some spices and turning it into sausage.
But the short players like Goldman set out to create sausage from pigs known to be sick because that would be more profitable for them, and this was a zero sum game: their profit came at the expense of their customers. Note that this is NOT inherent to investing, that the dealer’s gain is necessarily the customer’s loss. A dealer might exit a trade that he sees as unprofitable because he expect the price to fall in the next few days. The customer may have a completely different time horizon, and the success of his investment will not be affected much by what would be for him trading “noise” over the next few days.
Let’s put it more simply: how many of you would knowingly choose to be on the other side of a Goldman prop trade, particularly if you knew Goldman had designed the instrument to enable it to go short? Answer: probably zero.
There is an (in theory) less culpable scenario, but it does not get Goldman out of the SEC’s crosshairs. The initial motivation for its Abacus program (25 synthetic CDOs in total) was to lay off long CDS positions it took. Let us say a hedger like a bank wanted to reduce its subprime exposure. It could sell the loans or bonds, or simply hedge it by entering into a CDS with Goldman. From time to time, Goldman would flatten its position by bundling these exposures into a synthetic CDO. This was hardly unusual; a similar process was well established in the corporate CDS market.
So if that is the case (big if, one will have to look at Goldman’s intent, as revealed by internal messages, as to whether it was merely laying off exposures in a routine manner or cherry picking particularly drecky exposures to establish a profitable short), Goldman’s “we’re just acting as a market maker” argument is not a complete fabrication. But it still appears to have a legal problem. See this statement in its marketing documents (p. 346 of the Goldman documents released by the Senate):
Goldman Sachs has aligned incentives with the Hudson program by investing in a portion of equity and playing the ongoing role of Liquidation Agent.
Yves here. This is a flat out misrepresentation. The equity position is a Trojan horse for the much larger short position. The equity was at most 5% of an ABS CDO; the e-mails on preliminary deal structure show this one at 1% to 1.5%.; Goldman would be at least 98.5% net short this deal (if p. 401, which shows Goldman had a $8 million equity position, is correct, it was 99.6% short! And since per p. 402, it reported $17 million in P&L on the deal, so it took more out in fees than its equity stake. Nice work). It most certainly did NOT have incentives aligned with its investors
Goldman may argue that the disclaimer language in itty bitty print on the next page gets it off the hook, but I have my doubts that that will be viewed with much sympathy. There is a notion of good faith and fair dealing that underlies all contracts. It is such a bedrock concept that it is not clear that Goldman can try to disclaim its way out of it.
Goldman has more language that is misleading (p. 357):
Goldman Sachs’ objective is to develop a long term association with selected partners that can adapt to and take advantage of market opportunities
• The goal is to create attractive proprietary investments by leveraging expertise of both Goldman Sachs COO and Mortgage Desks while maintaining a consistent approach and creating a unified issuance program across multiple transactions
Yves here: Translation. We want to sell you more deals like this, so trust us, we won’t fleece you.
Note that Goldman explicitly says it is NOT laying off its own exposures, and by implication based on the body language thus far, it is pickin’ good stuff for this deal (p. 358):
• Goldman Sachs’ portfolio selection process:
• Assets sourced from the Street. Hudson Mezzanine Funding is not a Balance Sheet COO
• Goldman Sachs COO desk pre-screens and evaluates assets for portfolio suitability
• Goldman Sachs COO desk reviews individual assets in conjunction with respective mortgage trading
desks (Subprime , Midprime, Prime, etc.) and makes decision to add or decline
• All CDS use rating agency approved confirms (pay as you go)
It appears this deal was not an easy sale (p. 803, from an October 2006 e-mail by Michael Resnick):
do we have anything talking about how great the BBB sector of RMBS is at this point in time … a common response I am hearing on both Hudson’ HGSl 1s a concern about the housing market and BBB in particular!
We need to arm sales with a bit more – do we have anything?
Now Goldman defenders may argue that the investment bank is being unfairly singled out. However, that is hard to take seriously. There were very few banks in the business of synthetic CDO programs for their own account : Goldman, who is being investigated, Morgan Stanley, ditto, and Deutsche Bank….not. One industry source has also told us that Citigroup did deals along similar lines, but we have not gotten independent confirmation (update: aha, some new G2 in a very good article at the Financial Times tonight).
Some cynics may contend that the failure to go after Deutsche Bank is due to the fact that the head of SEC enforcement, Robert Khuzami, not only comes from Deutsche, but was involved in its CDO business. But the reality is more complicated. Getting someone like Khuzami, who is also a former prosecutor, was a coup for the SEC. He would clearly have to recuse himself from any cases involving his former employer. Insiders can correct me if I am wrong, but not only does the SEC not appear to have anyone who could step into Khuzami’s shoes (in terms of having both the product knowledge and the litigation experience), but it would be difficult to hire someone with a similar profile. Thus the fallback may be to perfect the litigation strategy on Goldman and Morgan so it then can then be deployed against Deutsche and not require someone as high powered to lead the effort.
Just because the wheels of justice seem to be grinding a bit slowly does not mean that in the end, they will not grind exceedingly fine.