Many of you asked for the BlackRock presentation that disclosed Goldman's desire to tear up AIG CDS contracts at a concession (i.e., discount), so here it is.
Some of the notable highlights first - as BlackRock points out on page 8 of the presentation:
- Negotiating Position: Goldman Sachs is the least risk averse counterparty, i.e., the only counterparty willing to tear up CDS with at agreed upon prices and retain CDO exposure
- Goldman approached AIG in August to discuss tearing up the CDS contracts
- BlackRock has advised AIG on tearing up 9 CDS in the Goldman portfolio with a $3 billion notional
- These transaction were selected because they were distressed portfion likely to experience credit events and convert to cash positions in the next few years (converting to cash position reduces any basis between the CDO and CDS values)
- The bid-offer spread between AIG and Goldman on these CDS tears-ups is $300 million
- Goldman has expressed a willingness to negotiate tear-ups on additional trades (including 7 synthetic Abacus transactions)
- Concession: Goldman would likely accept a small concession but may look to its funders to absorb the loss (or a portion of the loss)
- Goldman's exposure to AIG is limited to the difference between collateral requested (what they are likely posting to swap counterparties) and collateral received at any given time from AIG. While hedging this AIG counterparty risk is expensive, the cost would translate into no more than 2 points on the whole portfolio
- Goldman's swap counterparties are exposed to Goldman Sachs rather than AIG counterparty risk, and are therefore less likely to be receptive to deep concession
- Access to the Assets: Goldman has said that it does not hold the cash CDOs, but has back-to-back swaps on most of the positions
- Other factors
- Because Goldman prices have been consistently lower than third-party prices, Goldman and AIG have negotiated a collateral posting protocol in which Goldman's prices are given a 12% positive haircut for collateral posting
- Goldman's requested collateral margin, therefore, is generally 12% higher than the agreed collateral posting at any given time
This is why Goldman may have been willing to tear down contracts - in essence it would have had a loss buffer of up to12% more than prevailing market prices on the CDOs, so even a protection tear down would have resulted in the ability to take incremental losses on the underlying (assuming it was still held at AIG). Not only that, but Goldman's aggressive margin requirements were in a league of its own:
- $1.3 billion in additional collateral has been requested, but not posted as of 10/24
- Goldman's collateral request does not reflect any haircut to Goldman prices per the protocol established with AIG, so at least a 12% gap attributable to the haircut will remain as long as the haircut protocol is effective
- BlackRock's projected values are higher than collateral requested, i.e., BlackRock expects the portfolio to perform better than values implied by requested collateral
And here is why Goldman was particularly worried about the actual underlying securities in its portfolio:
The portfolio is projected to experience higher tranche principal losses than the overall portfolio in all cases (e.g., 15% higher than the total portfolio in the base case)
- Despite a significant concentration in prime/agency RMBS, the overall quality of the portfolio is impaired by a large exposure to Alt-A RMBS
- 29% of the Goldman portfolio is prime/agency securities concentrated in a few high-quality CDOs, compared to only 17% for the total portfolio
- However, another 26% of the portfolio is comprised of Alt-A RMBS (vs. 17% of the total)
- Additionally, 55% of Goldman's portfolio is concentrated in 2005 vintage assets (compared to 38% of total)
- By rating, Goldman's portfolio is barbelled
- 33% of assets rated AAA (compared to 36%), but 25% are below investment grade (compared 18%)
In layman's terms, what all this means, is that Goldman would have indeed been willing to accept tear downs due to the excess buffer (positive haircut) arrangement the firm had in place with AIG. Indeed, the firm had downside protection all the way down to at least 12% below fair value as determined by all other AIG counterparties (granted, in such an extremely illiquid market as CDOs nobody knew what price these securities would print at, especially if trades were done in the size discussed). Furthermore, according to BlackRock, Goldman was wise enough to offload the actual CDOs to clients, and was exposed merely through "back-to-back swaps on most of the positions." This means that Goldman only was exposed purely to counterparty risk on AIG's behalf - should AIG default, Goldman would become the client-facing entity guaranteeing "pass through" sold CDS.
Yet one wonders just how many billions of dollars Goldman had in margin variation between collateral posted to it via AIG, and how the amount of money it had paid to buyers of CDS sold by Goldman. We are certain that since no Goldman client had the same negotiating power as Goldman did with AIG, Goldman likely had a positive balance in the hundreds of millions if not billions simply on the collateral variance.
As page 10 indicates, whereas all counteparties had requested collateral at a price for the underlying CDOs of 49, Goldman was extremely aggressive, demanding collateral for a price-equivalent of 37. The latter compares to a BlackRock model price for Goldman of 44, meaning that even the Fed's advisor in good faith could not recommend such a generous treatment of Goldman in the context of all its other counterparties. Were Goldman to receive the same collateral as everyone else, it would be due 8.1 billion: $1 billion less than the 10/24 collateral request.
Now keep in mind, that of the top 5 counterparties, SocGen, GS, Deutsche Bank, Merrill and Calyon, only Goldman had subsequently sold off its entire CDO book: once again implying that unlike the other 4 firms, who at least held the exposure on their own balance sheet, and thus one can say deserved to receive some insurance, Goldman had bought then sold its entire portfolio, in essence making Goldman nothing than an AIG conduit, which was fully hedged and, as noted above, only had counterparty risk, yet which had the benefit of up to $1 billion in excess cash on its books due to day-to-day marking of its CDS exposure and its advantage collateral arrangement.
Futhermore, as Goldman owned CDS on AIG itself (as a counterparty hedge), Goldman had absolutely no risk in its relationship with AIG whatsoever!!! Of course, It is critical to remember that Goldman not only received par between the collateral previously posted to it, and actual cash from Maiden Lane III, it also made billions by selling actual AIG CDS (which as we claimed previously was done while in possession of material non-public information). Amusingly, while Goldman bought all of its CDO protection from AIG exclusively, it definitely used a very broad seller base when it loaded up on the actual AIG protection. Therefore, Goldman's only, ONLY risk, was that of a complete systemic collapse and the repudiation of all contracts, CDS and otherwise. Which is why Goldman's various agents did everything they could to prevent that from happening, up to and including loading the Federal Reserve with trillions of toxic debt in the upcoming 12 months.
And speaking of the Federal Reserve, while reading the SocGen section, we came across this little stuner on page 3:
We have heard second-hand from a trader that Soc Gen has pledged much of the portfolio to the Fed discount window for future liquidity
Aside from the fact that in October 2008 France-based Soc Gen was not a Primary Dealer (it only just applied for this position a few weeks ago), one needs to turn to page 5 of the presentation to realize that Soc Gen's portfolio had a value of 49 cents on the dollar. What this implies is that in October of last year (and ostensibly prior) Soc Gen, a foreign, non Primary Dealer, had access to the Federal Reserve's Discount Window, where it had pledged securities that had a value of 49 cents on the dollar, and for which the Fed would have taken arguably no haircut, thereby funding the French firm at par for securities that were worth less than half, and which the taxpayer was on the hook for. Indeed, these securities may well have been completely worthless: lower on page 3 we read:
Soc Gen and AIG are currently in dispute over existing events of default and credit events under transaction [ineligible] for 2 deals, totaling $650 million of notional exposure.
We would not be at all surprised if the defaulted securities were part of the crap that had been given to Tim Geithner, at the time head of the New York Federal Reserve.
And what Soc Gen was doing by pledging reference assets to the Fed, we are certain that all the other counterparties (those which unlike Goldman still held on to the securities) were doing as well.
The fact that the Fed was willing to risk taxpayer capital with such reckless abandon, first in the form of accepting literally worthless reference securities from Soc Gen (as documented by BlackRock), and subsequently by bailing out Goldman at well over par (remember the money the firm made on its actual AIG counterparty-risk) protection, would have been sufficient to terminate Geithner's career in any self-respecting banana republic. Too bad America is no longer even that.