Bankruptcy

Explaining the Abacus CDO

By Adam Levitin

I'm taking particular glee in the SEC's suit against Goldman Sachs for alleged market manipulation with the Abacus CDO.  This isn't Schadenfreude.  Instead, it's that I just taught synthetic CDOs to my structured finance class last week.  The timing couldn't have been better.  When was the last time a synthetic CDO made the front page of the Times?

The news accounts of the suit don't really do justice to the transaction at issue.  Hopefully this description helps a bit. 

In a nutshell, a synthetic CDO is a securitization of a portfolio of credit default swap positions. 

To explain more fully: 

A CDO is more or less a hedge fund.  It's an actively managed, unregulated investment fund.  The assets can be anything.  In the case of the Abacus 2007-ACA deal, the assets were a portfolio of credit default swaps (CDS).  The Abacus CDO was the securitization of a bunch of CDS positions (if it has cash flow, it can be securitized). 

The Abacus CDO was selling CDS protection on a bunch of dodgy mortgage-backed securities.   As long as there was not a credit event (however defined) with the MBS, the protection buyer (John Paulson) would pay protection premiums to the CDO.  If a credit event occurred, however, the payment flow would reverse, and the CDO would pay Paulson.  Presumably in this deal, Paulson did not own the MBS; he was using the CDS to take a short position on the MBS, with the CDO (and ultimately its investors) taking the long position.  

Goldman, the SEC alleges, constructed the CDO as a bespoke deal for Paulson, so there would be someone taking the long position against his short.  Goldman took a nice fee for this and is alleged to also have taken a short position (I'm not clear at this point if it was short on the MBS or if it bought CDS protection on the Abacus CDO, thereby taking a short position on the CDO, which is a derivative bet on Paulson's short). 

There are a lot of acronyms floating around here.  It's easier to understand the alleged fraud by analogy.  Think of Goldman as a fight promoter (Don King, as it were).  The promoter arranges the fight.  If the promoter is betting on the fight, there's a potential problem, though, as the promoter has the ability to fix the fight.   This basic fraud routine goes back to time immemorial.  The promoter arranges a fight between a big guy and a wimp.  Everyone bets on the big guy, except the promoter.  The promoter tells the big palooka that he's going down in the 6th round (and maybe gives him a cut).  The marks in the crowd get fleeced.  

The fight analogy also shows very clearly what the problem is.  The arranger of a fight (or deal) shouldn't be betting on it. 

Of course, then we have the Magnatar case.  So far there are no allegations that the deals' arranger was shorting the deals.  But the arranger still went along with creating a vehicle for a short that was not disclosed to investors.  I see the underwriter/arranger's role in the Magnatar deals as just as problematic as the hedge fund's--no arranger should be setting up a deal filled with what it knows are rotten assets.  Of course, as long as the arranger gets its fee, no questions asked.  This strikes me as a possible case of  aiding and abetting securities fraud.  But after Stonebridge, it's up to the SEC to bring such actions. 

All of this points to the problem with the PSLRA.  Congress thought the tort attorneys were getting out of hand.  And some were.  But as long as there was a real threat of securities litigation, it meant that attorneys (general counsel's office and outside counsel) played a self-regulatory role in the securities industry.  Private securities litigation thus had an important regulatory function.  The SEC can only do so much itself; Congress might have gone too far in curbing private securities litigation. 

Full post as published by Credit Slips on April 16, 2010 (boomark / email).