Goldman is a law unto itself, with no respect for propriety or notions like “fair dealing” (which in many contexts supercede particular contractual provisions).
The latest incident of Goldman scumminess is in one of the opaque corners of the credit markets, namely, collateralized debt obligations. Goldman launched a major program of synthetic collateralized debt obligations from 2006 into 2007. The early deals served to lay off its subprime risk; its later deals were mainly to dump CMBS risk.
The latest Goldman predatory practice is using contractual details (the deal equivalent of “gotchas” in credit card agreements) to undermine the well understood premise underlying ALL structured credits, namely, that bottom tranches take losses first. With two Abacus deals (and since these are Goldman programs, one has to assume it very knowingly planted these trap doors), Goldman is favoring the junior investors over the seniors. From Bloomberg:
Goldman Sachs Group Inc. paid off at face value some junior-ranking slices of two collateralized debt obligations at the potential expense of more-senior classes that now are likely to default, according to Fitch Ratings.
Goldman Sachs, the most-profitable securities firm, applied its “sole discretion” to ignore standard payment priority and use cash in reserve accounts for the Abacus 2006-13 and Abacus 2006-17 CDOs to retire lower-ranked notes, Fitch said yesterday in separate statements.
The moves are unusual in that the most senior creditors are typically the first in line to get paid. Fitch analyst Karen Trebach said the use of reserve funds may help cause or add to losses for holders of the CDO’s remaining classes.
“We are not aware of the use of this feature in other transactions we rate,” Trebach said in a telephone interview.
Yves here. This is the cute part, Goldman is using the cash in the CDO to redeem junior tranches that are at least impaired, and probably toast:
Goldman Sachs bet against $1.4 billion in commercial-backed securities and other CDOs through default swaps that were inserted into the Abacus issues, and it then sold notes from the CDOs equal to part of that amount, Fitch said. The cash raised was put in an “eligible investment account.” As with similar “synthetic” deals, securities bought for that account could be used to repay the CDOs, make payouts on the underlying swaps or both partially.
Goldman Sachs also had the ability to use “principal proceeds from the eligible investment account” to redeem Abacus classes “without regard to sequential order,” which it chose to do to retire junior classes, Fitch said.
Motivations for such action could include ownership of the notes or separate bets against higher classes, according to Howard Hill, a former Babson Capital Management LLC portfolio manager who founded securitization-related departments at four of the primary dealers that trade with the Federal Reserve, among them Deutsche Bank AG and UBS AG.
“You just don’t know without seeing who owns all the positions related to the deal,” Hill, who now runs a blog from New Milford, Connecticut, said in a telephone interview.
The reserve account may also be depleted because, while a Goldman Sachs unit entered into an agreement to buy securities in it at par to enable payouts if needed, that’s not true in all situations, according to Fitch.
The “put option” can be voided if any of those securities default, which is now likely as 38 percent of Abacus 2006-17’s investments are rated CC, Fitch said. Investments, which had to be AAA rated, may have included notes such as mortgage securities or other CDOs, Trebach said.
The “probable” default of account investments may also trigger the unwinding of the Abacus CDOs, and the amount garnered in sales may not be enough to terminate the underlying default swaps as would also be required, Fitch said. About 98 percent of swaps held by Abacus 2006-17 are tied to bonds with “junk” ratings or near that level and under review, the firm said.
If Goldman Sachs hadn’t used some of the cash in the accounts to redeem notes, the senior securities would be “less exposed to losses,” Trebach said. If what’s left is insufficient, even the Abacus CDOs’ super-senior swap classes may be called upon to make up the difference, she said.
These transactions also depended on credit enhancement, typically from an AAA guarantor, and an industry participant says that while the monoline ACA (a preferred dumping ground, and the lone non AAA rated monoline by design, which was rewarded for its credulousness by getting downgraded in December 2007 to CCC), it is likely pretty much all the logical suspects provided guarantees to various transactions in this program, including AIG:
I doubt that Goldman faced ACA direct on the Abacus deals, and there were prob a bunch of them done with MBIA, Ambac, and AIG. There is still TONS of CMBS resecuritizations out there where the monoline was the bagholder on the super senior (e.g. the later Abacus deals).
Why is AIG as possible AAA guarantor of interest? Because if Goldman was left holding any of the AAA tranches (not impossible at all, it might not have placed all of the trade), it was already taken care of via AIG. It might have bought the junior tranches at distressed prices and decided to self-deal. Another possibility is that some of the junior tranche holders are preferred customers (a lot of hedge funds engaged in correlations trades using the lower-rated CDO tranches). A former monoline executive speculates:
….these “pure synthetic deals” were different from the get go: they were never intended to be real bond offerings and of they were shorting the collateral, they wouldn’t have cared about senior class interests.
Is this related to the double payment issue?
If goldman gold made whole for their AIG exxposure via the government, perhaps they could not (or felt it wrong?) to collect again on the same bonds do they paid cashflow to the junior bonds?
The reason I doubt the benign interpretation, that Goldman was avoiding double-dipping, was that the firm is so busy trying to burnish its image that it would have made a point of this issue. Unless, of course, it does not want to remind the serfs that Goldman’s widely touted subprime short came at the expense of the great unwashed public.
Update 11/14, 12:00 AM: This is from someone who saw the deal docs. His remarks suggest disclosure might have been inadequate:
I have read hundreds of securtization disclosure documents (and drafted quite a few) and dozens of cdo dox. I have never seen any sort of “sub bond cross over date” that would apply for a deal that was taking losses. This would be such an unusual feature it would need to be highlighted in red and underlined, if I am understanding the facts in this article.
I have seen structures that deliberately fast paid sub bonds via excess spread (ie not locked out from principal for the first 3 years). But this was permitted because it did not harm the senior notes – ie sub bond was replaced by overcollateralization via excess spread (it also lowered the coat of the liability structure… And allowed bbb holders to get out ahead of senior holders). And it was always clearly disclosed and well discussed and understood.
This is happened on a 3 year old, under performing deal, so excess spread is not likely available.
I can’t think of a situation where a manager would get this type of discretion – most deals are written to protect against this type of “discretion” because it creates uncertainty about what a bond is likely to get. Changes in priority or cashflow ia always set up as something that happens by operation of triggers or events clearly labeled in the dox.
Mbs historically permitted a limited amount of discretion to manage defaulted loans (modify, short sale, pursue deficiency) which had some opportunity for conflict between them and the bond holders. As a result, there would be many checks in the dox to limit conflicts, such as requiring that the servicer not be a holder of the senior bonds…
I can’t figure this issue out. I know the abacus deals (and deutsche’s Start deals) – i reviewed them briefly before turning them down (the sales men were relentless on these deals). I saw enough to know they were different from normal transactions. All of the normal rules were off on the sales process, structuring, review etc.
Despite that, in reviewing the offering term sheets, I didn’t see broad cash flow priority discretion left to the manager. I would have been very surprised if I had.