If Goldman Sachs wanted to reduce its exposure to subprime mortgage investments, why didn't it simply sell the assets it owned? Two reasons:
First, those large sales would have sent a signal that something was terribly, terribly wrong, and thereby pushed prices down further. That's how supply and demand normally works.
Second, Goldman professed to be market maker, which uses its trading book to instill confidence. It ostensibly bought, sold and inventoried mortgage securities to provide stability and liquidity to the marketplace. Of course, we now know that such market confidence was entirely misplaced.
To sidestep these issues, Goldman and other major banks found a solution that subverted the laws of supply and demand, and escaped the price discovery of a transparent marketplace. They fabricated synthetic CDOs, such as Abacus 2007 AC-1. These toxic assets, invented out of thin air, made the meltdown worse than it otherwise would have been.
How much worse? Consider the numbers: According to the New York Fed, about $1.275 trillion in subprime mortgage-backed bonds were issued between 2004 and 2006. Of that amount, about $51 billion was rated in the BBB range. (In virtually every subprime bond deal, about 3% to 4% of the total capital structure allocated to tranches rated BBB+, BBB, and BBB-. Four percent of $1.275 trillion is $51 billion.)
So if Wall Street wanted to trade BBB rated subprime bonds, it had no more than $51 billion to work with. Some of those bonds were sold outright; others were repackaged into CDOs. The unfortunate investors who acquired those low-rated bonds, in whatever guise, could lose no more than $51 billion outright.
So how many mezzanine CDOs--i.e. CDOs stuffed which tranches rated BBB--were actually issued? Between May 2006 and April 2007, about $107 billion in mezzanine CDOs were issued, more than twice the physical supply of lower-rated bonds.
Wall Street transcended the limitations of physical supply via naked credit default swaps, which are like insurance policies on assets you never own. And a synthetic CDO may be nothing more than an assemblage of naked credit default swaps for cash settlement, meaning the reference mortgage bonds need never be physically delivered when the beneficiary gets paid. These deals were not financing anyone's home, directly or indirectly. Their failure cannot be blamed on the real estate bubble or irrational exuberance. As for hedging anyone's risk, the scant evidence available suggests that, at the time of closing, the fix was in. The probability of default had morphed from "highly likely" to "certain." Abacus 2007 AC-1 and its ilk served no legitimate business purpose.
Apparently, Fabrice Tourre saw Goldman's scheme for what it was. His emails read like a French epistolary novel, Les Courriels Dangereux, which tells a tale of love and betrayal amid the moral rot of the (financial) aristocracy. He writes:
"When I think that I had some input into the creation of this product (which by the way is a product of pure intellectual masturbation, the type of thing which you invent telling yourself: 'Well, what if we created a "thing", which has no purpose, which is absolutely conceptual and highly theoretical and which nobody knows how to price?') it sickens the heart to see it shot down in mid-flight... It's a little like Frankenstein turning against his own inventor ;)"
Even the Wall Street expression "IBG YBG" (I'll be gone, you'll be gone) has a French antecedent: Apres moi, le deluge.