Janet Tavakoli has always been more bark than bite. Being provocative is part of her shtick.
In her latest commentary, she accuses Goldman CFO David Viniar of lying about Goldman's exposure to AIG on a September 16, 2008 earnings call (the AIG bailout was negotiated later that day). This ridiculous conspiracy about Goldman and AIG just won't die, apparently.
On the call, Viniar said: "I would expect the direct impact of our credit exposure to [AIG] to be immaterial to our results." As Tavakoli acknowledges, it's a strong statement to say that a CFO lied to the public. It's also a patently absurd statement in this case. Yes, I know, Goldman is evil, Goldman owns the government, yada yada yada. Anyway, back in the real world, Goldman's exposure to AIG almost certainly was immaterial.
Let's go over this again. The total notional amount of CDS protection that Goldman bought from AIG was roughly $20 billion. But "exposure" in credit derivatives is equal to the cost of replacing a credit derivative in the market, not the notional amount of the transaction. Think about it this way: if you buy a $300,000 homeowners' insurance policy on your house, and your insurer goes bust, you're not out $300,000. The cost to you is simply the cost of buying another insurance policy to replace the first one. In Goldman's case, the cost of replacing its trades with AIG was about $10 billion. Against that $10 billion, Goldman held $7.5 billion in cash collateral. It then hedged the remaining $2.5 billion of exposure with CDS on AIG. This is why Viniar said that Goldman's direct exposure to AIG was immaterial.
So what are Tavakoli's arguments? One is the Immaculate Negotiation argument:
The government could have stepped in and renegotiated its contracts. ... Goldman Sachs would have been out billions of dollars in collateral had a bankruptcy‐like settlement been negotiated with AIG, and that is material.Saying that Goldman would've taken a material loss if "a bankruptcy‐like settlement been negotiated with AIG" is the equivalent of saying that Goldman would've taken a material loss if they'd agreed to take a material loss. It's true, but there's no way Goldman would ever have agreed to a "bankruptcy-like settlement" — why would they? As someone who has actually been involved in these kinds of negotiations, let me explain how the AIG/Goldman negotiations would have played out:
AIG: Would you be willing to accept, say, 70 cents on the dollar?Seriously, what could AIG have threatened Goldman with? If they didn't accept a haircut, AIG would file for bankruptcy? Fine, Goldman would've just seized the $7.5 billion in cash collateral, and collected the remaining $2.5 billion from its counterparties on the now-triggered CDS on AIG (on which more below), covering Goldman's full bilateral exposure to AIG. That's what it means to be "hedged."
(This is also why the Fed paid Goldman and the other counterparties 100 cents on the dollar to terminate their CDS contracts with AIG, which this Bloomberg article portrays as some sort of gift to the banks. But the Bloomberg article also relies on the Immaculate Negotiation argument — how, exactly, was the Fed supposed to get the counterparties to agree to take a haircut? The Fed had just demonstrated to the entire world that it wasn't willing to let AIG file for Chapter 11. How do you suppose those negotiations would have gone? The Fed couldn't say, "You can either take a haircut to 70 cents or AIG will file for bankruptcy and you'll only get 50 cents," because everyone knew the Fed wasn't willing to put AIG in bankruptcy.)
Now, with regard to that $2.5 billion in CDS on AIG, Tavakoli argues that "It is never a given that hedges will pay off when the chips are down." That's true, and there's no guarantee that the counterparties who sold CDS on AIG to Goldman would've been able to make the payouts. But these CDS trades, like most standard single-name CDS trades, were margined daily. At the close on September 16, the CDS spread on AIG was 53 percent upfront and 500bps running, which means that the counterparties who sold Goldman CDS on AIG would have already posted around $1.4 million in collateral (excluding Independent Amounts). So the maximum potential shortfall to Goldman was about $1.1 billion — and the only way they'd lose that amount is if the counterparties they bought CDS on AIG from all somehow couldn't pay. Viniar was entirely justified in assuming that these counterparties would've paid the remaining $1.1 billion.
Tavakoli also argues that Goldman had exposure because AIG's failure would have caused a system-wide meltdown:
If AIG had gone under, the already illiquid market would have frozen. Collateral requirements for all trading would have increased (just as they did the week Bear imploded), and Goldman would have had problems collecting from many trading counterparties.That may be true, but not only is that not the issue, there's also no way Viniar could possibly have known how the mayhem following an AIG default would have affected Goldman. It's not like he could've said, "Yes, our exposure is material because an AIG default will cause hedge fund clients A, B, and C to withdraw their prime brokerage accounts, initial margins to rise by X, and 2-year swap spreads to fall Y basis points." No one knew what would happen if AIG was allowed to file for Chapter 11 — not even the Great Janet Tavakoli. Remember also that what Viniar said was that he expected "the direct impact of [Goldman's] credit exposure" to be immaterial. He wasn't talking about Goldman's exposure to a broad systemic meltdown, and no one thought he was either.
Finally, Tavakoli argues that Goldman's exposure to AIG included "reputation risk." Yes, I'm sure that if AIG had failed, Goldman's reputation for having prudently managed its counterparty risk would've been devastating.
I own all 4 of Tavakoli's books, and it's undeniable that she's extremely smart. But like I said before, in her public commentary, she's all bark and no bite.