Inside the Goldman Sachs scam

May 5, 2010

Petrino DiLeo explains how Goldman Sachs set up its clients for a fall--in order to help a hedge fund manager who raked in $1 billion off their misfortune.

IN NOVEMBER 2009, Goldman Sachs CEO Lloyd Blankfein told a reporter (presumably with a straight face) that his firm was doing "God's work." He added: "We help companies to grow by helping them to raise capital. Companies that grow create wealth. This, in turn, allows people to have jobs that create more growth and more wealth. We have a social purpose."

But in the past few weeks, the world has gotten a clearer view of the work carried out by Goldman Sachs, a company aptly described by journalist Matt Taibbi as a "giant vampire squid wrapped around the face of humanity."

That work has nothing to do with God, social purpose or creating wealth for anybody other than Goldman and a few of its choicest clients.

The Securities and Exchange Commission's (SEC) fraud case against Goldman boils down to this: The firm is accused of conspiring with hedge fund boss John Paulson to design an investment (or a financial instrument, to use Wall Street-speak) called Abacus 2007-AC1.

This "instrument" was based on the most toxic residential mortgages that Paulson could put together. It was created so that a series of patsies would invest in these securities without knowing that Paulson was "shorting"--that is, betting against--the whole thing.

Lloyd Blankfein, CEO of Goldman Sachs, with James Dimon of JPMorgan Chase giving congressional testimony
Lloyd Blankfein, CEO of Goldman Sachs, with James Dimon of JPMorgan Chase giving congressional testimony

In other words, when Abacus soured, the investors got hammered--but Paulson scored, big time.

The Abacus deal was put together in early 2007, as it was becoming clear that the ever-expanding housing bubble was starting to leak. The deal was pushed through quickly because Paulson and Goldman realized it would soon be obvious to all that the bubble had popped--and that no one would be willing to invest in an instrument made up of so many risky mortgages.

Paulson's aim was to "short" the Abacus security through the use of credit-default swaps (CDS).

A CDS is supposed to be a kind of insurance, in which an investor pays a third party for financial protection. If the investment goes bad--if, for instance, bonds based on risky mortgages can't deliver their promised payout because the mortgagees defaulted--then the seller of the CDS is on the hook to pay the debt.

However, because CDSs aren't regulated, people can buy protection on assets they don't own. These are called a naked CDS. If naked CDSs sound like outright gambling, that's because they are.

This was Paulson's big bet. He loaded up on naked CDSs to cover Abacus 2007-AC1--the very financial instrument that he designed to fail. The companies that sold the CDS to him therefore paid out to him as the mortgages underlying the Abacus instrument went bad.

To pull this off, Goldman needed another firm--ACA Capital Holdings--to slap its seal of approval on the Abacus deal while keeping Paulson's intentions quiet. In turn, ACA's involvement gave Abacus the legitimacy necessary to convince yet other players, including big banks, to invest.

In the process, Paulson made a cool $1 billion, and Goldman earned massive fees for arranging the whole thing.

THE KEY piece to this whole charade is securitization--a process through which ordinary debts from credit cards, auto loans or mortgages are bundled together into huge securities that the biggest investors buy and sell.

These pools of debt are divided into "tranches" (the word's original meaning in French is "slice"). The various tranches pay out different levels of interest depending on the riskiness of the tranches. The riskiest bits pay out the most, but also suffer the first losses if the underlying debts--in this case, residential mortgages--go bad.

Through securitization, banks are able to move loans off their books and therefore rapidly expand their business. Rather than remain the creditor for mortgage holders for 30 years, the banks sell the mortgages to investment banks within weeks or months. For their part, the investment banks warehouse the loans, building them up into pools big enough to offer as giant securities--for which they down huge fees when deals for the securities are put together.

When housing prices were rising, investors couldn't get enough of mortgage-backed securities. There was literally no end to the demand from Wall Street for new mortgages that could be used to build new bonds.

It was this demand that led to the riskiest mortgages proliferating. One example is stated-income mortgages--where borrowers weren't required to provide any proof of their income to qualify. Still dicier were so-called NINJA loans, the industry's cynical nickname for borrowers with "no income, no job, no assets."

But even this incessant production of mortgages wasn't enough to feed the hunger on Wall Street. That's why the financial wizards developed collateralized debt obligations--or CDOs.

These instruments belong to a broad category of financial instruments known as derivatives because their value is "derived" from the price of an underlying asset. CDOs, for example, aren't created directly from mortgages. Instead, they are the repackaged assets of other securitizations, such as CDSs.

The Abacus deal at the center of the SEC lawsuit against Goldman was composed of the riskiest type of derivative--a synthetic CDO. These CDOs are called "synthetic" because they have no direct link to mortgages. Essentially, they were bets by bystanders on whether or not mortgages contained in other bonds would pay off.

The Abacus deals were synthetic CDOs tied to mostly sub-prime home loans and commercial mortgages. And because they were so profitable, Goldman Sachs had plenty of competition in churning out CDOs--for example, from Switzerland's UBS and Germany's Deutsche Bank.

As Roger Lowenstein described the Abacus deal in a New York Times column: "[T]he investors weren't truly 'investing,' they were gambling on the success or failure of the bonds that actually did own mortgages. Some parties bet that the mortgage bonds would pay off; others (notably the hedge fund manager John Paulson) bet that they would fail. But no actual bonds and no actual mortgages were created or owned by the parties involved."

Many CDOs are so complex that they become "effectively unanalyzable," according to financial blogger Yves Smith.

For example, an asset-backed security CDO could include 100 to 250 financial instruments. Each instrument, in turn, could be built of around 5,000 mortgages. As a result, a purchaser of one CDO could be exposed to the fortunes of 400,000 to 1 million mortgages without actually owning any of the loans.

But it gets even more insane, because a portion of one CDO may include tranches of other CDOs.

It was this interlocking web of CDOs, like the Abacus deal, that ultimately crippled the financial system in late 2008. When the housing market started going down and mortgage defaults proliferated, this poisoned countless CDOs--many of which were thought to be unconnected to the housing market. With no bank sure of how other banks were exposed or what they held, confidence in the financial system was lost.

THIS PRACTICE was at the heart of Goldman's Abacus deal. It was a synthetic CDO derived from existing residential mortgage-backed securities. Paulson designed a CDO tied to the most overpriced housing market.

Paulson's scheme was to buy a CDS from Goldman. He would pay a one-time fee, and if he was right about the CDS paying off, Goldman would pay him up to $1 billion. But Goldman needed a way to cover the bet, so Paulson suggested building the CDO and finding a way to convince investors that it was a sound deal.

That's where Goldman Vice President Fabrice Tourre, who is named in the SEC's suit, came into the picture. Tourre was charged with creating the Abacus CDO and approaching ACA to help. Ultimately, ACA picked 55 of Paulson's securities to include in the CDO. Tourre then marketed the CDO with no mention of Paulson or his bet against the assets. According to the SEC, this constitutes fraud.

The Abacus CDO closed in April 2007. But by October of that year, 83 percent of the securities it contained had been downgraded. By January 2008, 99 percent had been downgraded and lost substantial value.

The collapse of Abacus meant that Paulson's CDS soared in value and netted $1 billion. Goldman, in turn, got millions from investors. Dutch bank ABN/AMRO alone paid Goldman $841 million in connection with the deal--and much of that money went to Paulson.

The Abacus deal was hardly an anomaly. It's more likely the tip of the iceberg, as the recent ProPublica exposé of hedge fund Magnetar illustrates, creating doomed CDOs was widespread on Wall Street.

As Nomi Prins, a former managing director at Goldman Sachs, put it, "[B]anking businesses that are tied to the real economy are dying, but raw gambling disguised as finance is doing fine. Wall Street is making money by rolling the dice---again. All this risky activity seems to be going unnoticed in Washington."

Further Reading

From the archives