Did Goldman Sell Its $2.5 Billion AIG CDS While In Possession Of Material, Non-Public Information?

As frequent readers will recall, on March 20, 2009, David Viniar and Lucas van Praag held a conference call in which, for the first time, the firm discussed the impact of its hedging on AIG as well as its collateral and otherwise exposure with the nationalized insurer. The full transcript of the call can be found here, while our immediate comments on call, which we posted after the call, are replicated below in their entirety:

In a nutshell - Goldman had bought billions in AIG CDS in the 2004 to 2006 timeframe. Whether this was predicated by their expectation that subprime would blow up, or their very early understanding just how bad things at AIG were, one will never know, especially not the SEC. However, one look at the CDS chart below shows what prevailing levels for AIG's CDS was in that time frame. As one can see, AIG 5 yr CDS traded in a range of 4 bps to 52.50 bps between October 1, 2004 (only goes back so far) and December 31, 2006. Indicatively 5 yr CDS closed yesterday at a comparable running spread equivalent of 1,942 bps.

Purchasing $10 billion in CDS (roughly in line with what Viniar claims happened) at a hypothetical average price of 25 bps (and realistically much less than that) and rolling that would imply that at today's AIG 5 yr CDS price of 1,942 bps, the company made roughly $4.7 billion in profit from shorting AIG alone! This would more than make up for the $2.5 billion collateral shortfall (out of $4.4 billion total) GS claims AIG had with Goldman Sachs... If AIG had filed for bankruptcy, and assuming Lehman is any indication, the P&L would have likely hit $6+ billion.

Implicitly, one could say GS was incentivized to see AIG fail. Does that maybe answer some of the questions of why GS allegedly pulled AIG's collateral and started the avalanche that lead to its bailout? However, a fine point - if AIG had really tanked none of the CDS would be collectible as the entire CDS market would have likely imploded... Thus demonstrating the need for a zombie bank system: not totally dead (systemic collapse) but barely alive to pocket a nice little CDS annuity from daily cash collateral posts as it leaks wider (and taxpayers foot the bill).

Now that the dust has settled somewhat, we can refine this analysis: we now know that our original read of Viniar's words was excessive, and in fact Goldman only had $2.5 billion in long AIG CDS exposure. The linear relationship on P&L, keeping in line with the above math, means the profit on the CDS transaction was a quarter of the suggested $4.7 billion, or roughly $1.2 billion: still a staggering number for a firm that fundamentally did not experience any risk with regard to its CDO exposure, which as everyone and the kitchen sink now know, was paid down to goldman at par, thus assuring no P&L hit on the underlying security.

So far so good.

Yet the reason we bring up the topic of Goldman's AIG CDS transaction is due to a post James Keller which points out some distubring observations.

The bigger story, still unexplored by regulators and Goldman's critics, concerns that $2.5 billion in protection Goldman had acquired before the crisis hit. What became of those hedges? What did Goldman Sachs do with its AIG protection?

Goldman spokesman Lucas van Praag made a disingenuous case in a letter to the Wall Street Journal last April. He implied that nothing much became of the hedges: "In order to collect under a credit default swap, there has to be an event of default. No event of default means no payout." Goldman would have us believe that since AIG did not default, the CDSs expired and vanished forever.

The latter is of course not the case, and as we demonstrated 9 months ago, the company very likely sold its protection at a massive profit to cost. Keller continues:

This is not quite right. A company can avoid default, but one can make a lot of money selling protection on the company when everybody else thinks they are going to default. That is especially true if one has been involved in meetings with the Fed where the subject of the meetings was how to avoid such a default. A good trader buys protection when a company seems very safe, and sells it when the company seems very risky. It is not appropriate for a trader to sell protection when he has non-public information that the risks have diminished, that the government has unequivocally committed to saving this company.

The last sentence is critical, as while debate may rage over the definition of front running as it applies to Goldman, its clients, and its prop desk, trading on material, non-public information is broadly frowned upon by everyone in the marketplace.

When did Goldman sell its $2.5 billion of AIG protection? Goldman representatives have said that the protection was sold in the six months following the September 15, 2008 bailout loan. This is problematic. That is so because the details of the bailout were not released until March 15 of last year, when the famous AIG counterparty payments at last became public.

This suggests that Goldman sold its protection to counterparties that knew materially less about the actual risk of AIG than Goldman did. Remember that this bailout was specifically designed to avoid an AIG default, the event that forces Credit Default Swaps to be triggered. Goldman, in frequent conversations with Paulson and Geithner, knew that the government had just committed $85 billion to avoid exactly this outcome.

And it keeps on getting worse for the vampire squid.

The rest of the world, purposely kept in the dark, saw the risk of an AIG failure as imminent. In fact, on September 21, 2008 then-Treasury Secretary Paulson went on Meet the Press and explained that the $85 billion bailout loan would "allow the government to liquidate this company." Paulson may have been speaking loosely, but in the specific language of Credit Default Swaps a government "liquidation" is a Credit Event akin to bankruptcy and would "trigger" these swaps. This is why, immediately after the bailout was announced, the cost of protecting AIG risk skyrocketed. It rose to more than 40% of the amount hedged; meaning that Goldman, which had $2.5 billion in hedges, would have been sitting on over a $1 billion profit. [This number is in line with our estimates]

Not bad, even for Goldman. The government had just met their collateral calls, their risk to AIG was gone, and their hedges were in the money by $1 billion. What to do now? The right thing is to not sell the protection until the full details of the bailout are in the public domain. To wait until you have no material, non-public information.

It appears Goldman did not wait until after March to sell its protection. Yet Goldman has denied making a windfall gain on AIG. AIG CDS protection ended the second quarter of 2008 at about 200 basis points. From September 15, 2008 to March 15, 2009 AIG CDS never closed below 400 basis points. It is hard to see how Goldman could not have had a windfall. Depending on when Goldman sold this protection, the gain could have been as much as $1.5 billion. This would mean that while most of AIG's counterparties got 100 cents on the dollar, Goldman actually got far more.

This was the point of our original post all along, yet at that point we ignored the implications of the just released revelations of AIG's counteparties "make whole" agreements, which were certainly not public prior to March 15, except to Goldman and a select other few. The only exculpation is if such material CDS trades occurred in the presence of "big boy" letters, which as far as we know has never occurred.

In conclusion we thoroughly agree with Keller, that while anger at Geithner over the AIG fiasco is well warranted, the true culprit, with potential alleged elements of actual criminality even according to our flawed, corrupt and broken regulatory system, is and has always been Goldman Sachs.

Criticism of Geithner seems appropriate. Paying counterparties 100 cents on the dollar was unnecessary. Keeping the whole thing a secret was indefensible. Allowing windfall profits was unconscionable. But the Fed's behavior may not be the worst element of this episode.

Frankly, it is hard to see how, in having sold its AIG protection before March 2009, that Goldman Sachs can avoid the appearance that these trades were improper. Over a year on, we still await a clear explanation of how much the firm made from this protection and how its subsequent sale can be justified when Goldman had information that the federal government was deliberately keeping from the public.

Once again, we encourage Congress to invite not just Tim Geithner, but Messers Van Praag and David Viniar to provide additional disclosure on the exact timing of their AIG CDS sales, and specifically whether these transactions occurred prior to March 15, while the firm was in possession of material non-public information.

Update: For those who claim the SEC has no regulation over CDS trading, we strongly suggest you read the following.

While it is worth noting that the definition of ?security-based swap agreement? has never be litigated in this context (a point the defendants are likely to stress), the commission will likely argue that CDSs are security-based swaps subject to the SEC‘s antifraud jurisdiction. Under the GLBA, a ?swap agreement? is ?any individually negotiated contract, agreement, warrant, note, or option that is based, in whole or in part, on the value of, any interest in, or any quantitative measure or the occurrence of any event relating to, one or more commodities, securities, currencies, interest or other rates, indices or other assets ...[22]

In other words, ?an OTC derivative is "securities-related" when the reference is to an entity that is an issuer of securities (such as a public company), to a security itself (or a related event such as a dividend payment), to a group or index of securities or issuers, or based on related aspects of a security or group or index of securities or issuers, such as price, yield, volatility, dividend payments or value.?[23]

The SEC will likely argue that the VNU-related CDSs referenced bonds issued by the VNU holding company, and that such bonds are securities, therefore, the CDSs are "security-based."

The SEC will attempt to persuade the court to recognize that CDSs and the obligations they reference are so interrelated, that a finding against them would cause of anomalous result of permitting the SEC to pursue insider trading actions involving derivatives based on domestically traded securities but not those based on securities traded elsewhere, especially where derivative transactions might have effects in the United States. The commission will likely point to other ?long-arm? cases it has brought (and won) to convince the court that the defendants should be subjected to the commission‘s enforcement jurisdiction.