The ABC’s of Goldman’s ABACUSApr 29th, 2010 | By Rob | Category: Banking, DS Feature, Industry Analysis
Whether you’re an institutional investor or a stay at home mom who follows CNN, you’ve most likely heard the words “ABACUS”, “SEC”, “Goldman Sachs”, and “Fruad” in the same sentence. Most will probably tell you that ABACUS was a fancy deceptive ploy by Goldmand Sachs (GS) to cheat other banks out of their money and make bets against the U.S. housing market, which would ultimately lead to the recession. To those views we’d have to say “maybe”, “no”, “no”, and “definitely not”. Call us what you will, but we’re going to take a different stance on the issue, first explaining what ABACUS really is and then placing it in the context of financial markets.
Sure it’s easy to love a scandal when the home team (main street) is hurting, but smearing your rival’s leading guard (GS), because he continuously posts “tripple doubles”, doesn’t necessarily bring the trophy home. Goldman Sachs has been the source of an accelerating wave of controversy over the past few weeks, due to the SEC suit against the firm for fraud, and was most recently bent over the well worn U.S. Congressional hearing desk, compliments of U.S. Senator Carl Levin, of Michigan.
There’s a lot of chatter on the blogosphere about the actual construction of the ABACUS deal and just how it went down. Most of which requires time on an investment banking desk or a masters in quantitative finance to understand. We’re going to take a stab at making it one big step simpler…
What is a CDO?
If you don’t know this then you’re already lost. A CDO is a Collateralized Debt Obligation, which means that it is a security that is priced to a certain degree of risk based on a diversified pool of a particular type of debt assets. Generally CDO’s are set up by collecting a pool of cash flow producing assets, in the case of the recent crisis many CDO’s were actually Mortgage Backed Securities, thus the term MBS. The pool of cash flow producing assets is diversified across the nation and is then divided into levels of risk, known as “tranches”. The risk levels have nothing to do with the quality of any one mortgage or borrower. Instead the risk levels are simply a pecking order of who loses, from first to last, when payments from any asset in the pool stop coming in. See this excellent explanation of mortgage based CDO’s for a crash course.
What is a Synthetic CDO?
Now let’s try to get a grasp on synthetic CDO’s. If a normal CDO is tied to cash flow producing assets, a synthetic CDO is priced based on a portfolio of cash flow producing assets in a far more indirect manner. Synthetic CDO’s are derived by selling Credit Default Swaps (CDS) on cash flow producing assets (bonds, mortgages, etc.) in order to receive periodic interest payments that theoretically mimic the cash flows of the pool of performing loans. If CDS’s are the insurance on a pool of assets going belly up, then synthetic CDO’s are priced by extrapolating the present value of future cash flows coming out of the CDS’s of that pool. The synthetic CDO’s are then tranched from first to last loss, just like a normal CDO, where the riskiest tranche is always the highest yielding. The most common tranche categories from first loss to last, are “first-loss (FL)”, “D”, “C”, “B”, “A-2″, “A-1″, and “Super Senior (SS)”.
For Synthetic CDO’s to exist there must be a “market maker”, which is comparable to the “bookie” in a high stakes gambling circuit. These contracts and securities are generally formed and exchanged between few specific parties and organized by an investment bank, rather than a central clearing house (e.g. NYMEX, CBOE, etc.), and in the case of ABACUS the bank “making the market” was Goldman Sachs.
(Review note: FL will always be the first investor group to lose their principle, while SS will be the last. For SS to go belly up, the entire pool of assets tied to the CDS must default. When all loans are performing, CDS insurance premiums continue to roll in and Sythetic CDS’s yield large profits, mainly due to the small amount of principle required to sell the CDS’s and thus open a synthetic CDO position.)
What is ABACUS?
Abacus is a whole new breed. The idea behind this “business entity”, as it was described by its founders at GS, was to create two tranches of a synthetic CDO, which weren’t actually tied to any real cash flow producing assets at all. If a CDO is a fish in the hand, and a Synthetic CDO is a fish on the line, this baby cut the line and instead imagined the fish on the other end, basing their assumption of the fish on a bunch of other fish that had recently been caught.
Before we go any further, take a look at this illustration to get a sense of the way the ABACUS deal was structured. It’s not perfect, but it clearly shows that there are only two tranches in this mechanism, A-1 and A-2. If you are comfortable with the definition of a Synthetic CDO above, you should be asking, “how can you only have AAA and AA rated tranches if the lower tranches don’t go belly up first?” Isn’t the point of the tranches to create a pecking order of defaults? In a Synthetic CDO “yes”, in ABACUS “no”.
This is where the lines begin to blur… ABACUS Ltd. is specifically a business entity, organized and formed by Goldman for the purpose of market making. According to the SEC Goldman formed ABACUS on a request from John Paulson’s hedge fund, in order to make a market for Paulson to buy a quantity of AAA and AA rated Synthetic CDO tranches. In this way Paulson could purchase securities, designed to mimic CDS’s on AAA and AA tranches of a residential mortgage pool without having to find a buyer of the CDS’s of lower tiered tranches as with a normal or synthetic CDO contract.
The two lucky souls who ended up responding to the market making were IKB Deutche Industriebank and ACA Capital. Involved in insurance, corporate bonds, and municipal debt, these firms were clearly more than able to perform the due diligence to realize the risks of their positions, and were interested in placing a long bet on the strength of the U.S. housing market. By entering into the contract these two firms sold Paulson the securities designed to mimic AAA and AA CDS tranches, thus making themselves liable to pay the full principle of a the theoretical value of the reference portfolio in exchange for a handsome cash flow premium.
As we all know, synthetic CDO’s and assuredly agreements such as ABACUS performed very poorly for the sellers of the CDS contracts (e.g. the side long the housing market), when the reference portfolios crashed towards zero and for months no market was successfully made. Many similar agreements had been made and the volume of bets surrounding the housing market far surpassed the actual amount of mortgage debt issued, further muddying the market’s perception of what these reference portfolios were even worth.
What you should know is that Goldman Sachs was not along in this levered gambling of mortgage debt. Many parties came to the table wanting to get a piece of these long housing market positions (synthetic CDO’s and ABACUS style CDO’s) which generated cash upon issuance, like a short sale, and offered the potential for profit if reference portfolios continued to appreciate.
The blame here rests solely on the greed and ignorance of a whole industry bent on better than possible returns at smaller theoretical risk. Anyone who was a part of this investment banking market making process and had access to the volume of transactions being done, could have seen the excesses in their purest forms. Goldman was hired to make a market where a new type of security could be sold. Granted, the security was bogus, but there were no lies being peddled in the process and the prospectus clearly stated the terms upon which ACA and IKB must agree in order to have this theoretical CDO exposure. Whether Paulson’s fund had a hand in developing ABACUS or not, Goldman simply offered a product to consumers ad hoc and these two banks felt that the terms were fair enough to engage. They ended up making a terrible long bet on the U.S. housing market near the end of the party. They lost. End of Story