Goldman Wasn't Alone in Deals Like Abacus

Mark Wilson / Getty Images

"This is Lloyd on Sunday in New York," the voice mail began. Two days after the SEC sued Goldman Sachs for securities fraud, CEO Lloyd Blankfein left a message for his 30,000 employees, reminding them of the firm's core values: "teamwork, excellence, and service to our clients." He also tried to put into context the government's charges, which allege that Goldman failed to disclose key facts about a security it sold in 2007 called Abacus. "Importantly, we had assumed risk in the deal and we lost money," Blankfein said, "just like the other two long investors."

Goldman says it lost "in excess of $100 million" on Abacus, and may lose a lot more. The SEC is alleging that Goldman, whose near-mythic status on Wall Street has suffered a series of knocks, knew a critical detail the others did not: that hedge-fund manager John Paulson had helped pick the toxic assets that served as collateral for Abacus. Paulson had always planned to bet against the risky vehicle, and his fund made about $1 billion when Abacus blew up within a year of its creation. Two other investors, German bank IKB and Dutch bank ABN AMRO, together lost more than $1 billion. In other words, in exchange for a fee, Goldman allegedly assembled a house using subpar materials so one of its clients could bet on its collapse—and then sold the house to other customers.

Goldman says all investors knew exactly what was in Abacus. It has vowed to "vigorously defend" itself against the SEC charges and has hired former White House lawyer Gregory Craig. So it could be many months before we know who's right or if the bank broke the law. What we do know, however, is that as the housing market peaked in 2006 and 2007 several other big banks, including Merrill Lynch and UBS, did deals very similar to Abacus. The charges against Goldman provide a window into the nature of these arcane financial instruments, which were integral to Wall Street's meltdown.

Abacus is what's known as a synthetic collateralized debt obligation. A CDO is a financial tool that repackages individual loans into a product that can be chopped up, repackaged, and sold on the secondary market. They are "collateralized" in that they are backed by loans, bonds, or other real assets. As interest rates plummeted after 9/11, investors worldwide were eager for the cash flow being generated by millions of new American mortgages. Soon there weren't enough mortgage bonds to satisfy demand, so bankers hit on the idea of the synthetic CDO, basically a bundle of credit default swaps (or insurance contracts) that mimic, or reference, the performance of real bonds. By 2005, the CDO market in the U.S. hit $200 billion, twice the 2004 level. By then, housing prices were sky-high and the Fed had begun raising interest rates. A few smart investors believed the market was overheating and that CDO volume would drop.

Instead, it nearly doubled in 2006, to $386 billion in the U.S., and more than $520 billion worldwide, as banks grew more creative in the way they assembled and marketed the instruments. Some allowed favored clients, often hedge funds, to help choose risky assets, and then bet against the whole thing.

The Chicago-based hedge fund Magnetar Capital was allegedly among the first outfits to employ this strategy. Named for a type of neutron star that crushes anything that gets near it, the fund launched in spring 2005. Magnetar would agree to buy the riskiest piece of a CDO, which helped Wall Street firms draw in other investors. Next, according to an investigation by journalists at the nonprofit ProPublica, Magnetar pressed Wall Street firms to include junky bonds in the CDOs (Magnetar has denied this) so that it could bet against them. When the CDOs defaulted, Magnetar lost the small amount it had invested but made much more on its short bets. The strategy became known in the industry as the Magnetar Trade. From 2006 through 2007, Magnetar sponsored 30 CDOs, most of them synthetic, worth more than $40 billion. By the end of 2008, 95 percent were in default, according to ProPublica.

One of the deals Magnetar sponsored was called Norma, a $1.5 billion, mostly synthetic CDO that Merrill Lynch put together in March 2007. The CDO went bust within a year of its creation, and is now the subject of a lawsuit filed against Merrill Lynch by Rabobank, a Dutch firm that lost $45 million on Norma. The lawsuit alleges that Merrill Lynch did something very similar to what Goldman did with Abacus: it created a CDO designed at the behest of a hedge fund that wanted to bet against it but didn't disclose the role of the short seller to investors. Rabobank alleges that Norma was "a scheme employed by Merrill Lynch to maximize its own revenues and mask its losses during the last gasp of the subprime mortgage boom." Merrill Lynch says that it gave Rabobank "every piece of information requested and that was required."

Another lawsuit, filed by Connecticut hedge fund Pursuit Partners against UBS, makes slightly different allegations concerning a similar deal. Pursuit invested $35 million in multiple synthetic CDOs, which UBS sold from June to October 2007, days before the ratings agency Moodys announced a change in the way it would rate synthetic CDOs. This change led to an across-the-board downgrade. The hedge fund alleges that UBS learned of the pending change in a secret meeting with Moodys, yet failed to disclose it. The suit includes internal e-mails in which UBS employees refer to their CDOs as "vomit," and boast that they "sold more crap to Pursuit." In September 2009, a Connecticut judge ruled in favor of Pursuit, stating that the deal was akin to UBS's handing a gun to the plaintiffs and telling them it wasn't loaded, "when in fact UBS knew the gun was not only loaded, but was about to go off." UBS appealed the decision to the Connecticut State Supreme Court, and says that Pursuit was a sophisticated investor that knew it was buying troubled securities at deep discounts.

Banks and hedge funds weren't the only ones investing in these CDOs. King County, Wash., and the Iowa Student Loan Liquidity Corp. have both sued IKB, which also lost money in the Abacus deal, over a highly rated investment it put together in June 2007 called Rhinebridge. In what the suits allege was "the shortest-lived Triple-A fund in the history of corporate finance," Rhinebridge was downgraded to junk status within four months.

The consequences of these deals have rippled around the world. An estimated 50 million people were affected by the Goldman Abacus deal, including home-owners whose mortgages were part of the CDO, as well as institutional investors and shareholders of the foreign banks, according to After the Abacus deal went bad, ABN AMRO was bought by a consortium of banks led by the Royal Bank of Scotland, which in turn was taken over by the British government. British taxpayers now own 84 percent of the bank.

But do these deals represent fraud or just the cold nature of financial markets? All the parties claiming foul play are sophisticated players. And though the banks didn't in all cases disclose the role of short sellers in structuring the products, investors were aware of what was in each portfolio. "Do you blame the junkie or the dealer?" asks Rohan Douglas, who ran Salomon Brothers' and Citigroup's global-credit-swaps research divisions through the 1990s. In other words, Wall Street was selling a toxic substance to an addicted consumer. But in the U.S., at least, the law typically imposes stiffer sentences on those dealing drugs than on those buying them.

Join the Discussion