Goldman Under Fire

BACKGROUND OF THE DEAL

ABACUS 2007-AC1 was one of the over 20 synthetic CDO transactions that Goldman marketed to clients from 2004-2007. All of the Abacus deals were constructed in a similar manner using an SPV (Special Purpose Vehicle) to issue CLN’s (Credit Linked Notes) on a reference portfolio consisting of RMBS (Residential Mortgage Backed Securities). Notably the performance of the CLN’s was tied to a CDS (Credit Default Swap) on the same reference obligation. Later material will go more in depth as to how the particular products work, but in this Abacus deal, according to the SEC complaint prominent hedge fund Paulson & Company directly influenced the Portfolio Selection Agent’s decision – in this case ACA Management, LLC., to include mortgage securities that were more likely to suffer losses. The SEC has gathered a significant amount of evidence including emails and phone calls between Goldman, ACA, and Paulson & Company proving this point. Yet, as these allegations will be disputed in court very soon it would serve more prudent to discuss all the parties involved.

PLAYERS

Goldman in this instance served as the initial mediator that brought all the separate parties together. This included investors like IKB - a German commercial bank, CDS protection buyers – mainly hedge funds under Paulson & Co., and even several trading desks within Goldman. Investors would look to buy CLN’s on the reference portfolio that paid interest and returned principal at maturity, in the hope to gain loan portfolio exposure outside of making traditional options. This has been a very common practice for banks since the creation of the mortgage backed security in the 1970’s.

During this time as investors were looking for places to allocate capital Paulson & Co. was searching for potential investments and saw the subprime U.S. residential mortgage market as being on the verge of collapse.  Contrarian investing at its finest (and a bet that paid off in the billions) led the large fund to build short credit positions in CDS by buying protection on RMBS. This option was preferable because it is often very difficult to short bespoke, specific MBS; not the mention the high negative carry associated with the trade. At the same time, buying protection was available at extremely low cost considering CDO valuation models had very little historical data available and were skewed towards the recent outperformance of CDO tranches. Even in the ABACUS pitchbook, ACA asserted that their track record never included any notes within a CDO selection had been downgraded. This period was filled with euphoria and a broad overextension of credit that eventually led to strong shocks to the financial system.

Goldman, the last major player, had started to foresee some of the trouble for subprime markets according to several internal emails and memos, yet they were still reluctant to pass up revenue. They were involved in the initial transaction by first purchasing protection via swaps on all the A through D classes. At the same time they were responsible for providing the collateral for all the CLN and conducting a basis swap with the SPV (ABACUS). Though the only fee they claimed they received was from Paulson & Co. for $15 million they also had the opportunity to hold all the short credit positions and at this time it is unknown if all were transferred through its prime brokerage to hedge funds or some kept on their books.

PRODUCTS

Financial engineering has led to a variety of innovations including the many discussed here, though in reality many are simple in concept. In its most basic form the synthetic Collateralized Debt Obligation is a modification and reapplication of the banking business model of borrowing short and lending long. Investors sell protection through a Credit Linked Note (where the investor buys a note, that money is then used to purchase AAA rated collateral - in most cases U.S. Treasuries. Collateral is then stored in the case of a credit event where it is used to pay for the CDS protection buyer in full. A credit default swap is no different in its foundations than an insurance policy where protection is bought or sold on the underlying entity.), where in return their money is given to the Special Purpose Vehicle to both enter in swaps on their behalf and purchase collateral. ABACUS was created as an SPV and in the United States theses are often formed as trusts; this ensures little counterparty risk and replicates a micro-clearing structure. SPV’s are also able to issue debt which once again to repeat in this case are the CLN.

Quarterly payments are made by the protection buyers to the SPV which converts these payments into note coupons for investors. Since MBS are susceptible to prepayment risk, any return of principal on the reference obligation will result in the sequential amortization of CLN starting with the most senior tranche. Jumping back to CDO’s it is key to understand that how subordination of losses affects the structure and composition. CDO’s are built of tranches that have attachment and detachment points for losses (see picture). This enables investors to purchase a tranche that most replicates their investment needs. For example many insurance companies are typically involved in purchasing Supersenior debt that pays very little spread, but have the most subordination. However on the other end of spectrum, the first loss tranche or also referred to as the equity tranche can prove to be a valuable investment in the case that default correlation is very high (for traders) or realized default rates are low.

In the event or downgrade within a CDO the SPV passes the collateral to the protection buyer who is then able to sell it on the open market. At the same time, it returns the reference entity to the CLN investor who then attempts to recover its losses.

2007 DEFAULTS

Subprime mortgages collapsed just as Paulson had predicted, and within 6 months 83% of RMBS on which the CDO had been constructed had been downgraded. Within 9 months 99%, according the SEC’s research, had been downgraded. Investor losses amount to over $1 Billion which equated to nearly equivalent gains for Paulson’s hedge fund. What was interesting, and what investors like IKB had missed were that the tranches were rated at least A2 by Moody’s, yet the subprime bonds were all Baa2. At the same time the junior tranches (D-B, see pitchbook), were progressively higher rated, yet individual tranche widths were narrow enough that in any serious set of credit events - all would be wiped out.

9 months later that was exactly the outcome as default correlation across the assets exploded resulting in fat tail for loss distribution. Losses affected even the Supersenior tranches which are short default correlation, in the case of Goldman who held a $90 million position on its books. In reality the only good position was buying protection, as any trade involving purchasing equity positions (effectively long default correlation) to profit off Mark to Market gains, would have most likely unprofitable.

IMPLICATIONS

Paulson & Company was hailed for their massive profits betting on the subprime market crash from across the financial community. Little was known or spoken of at the time of how they orchestrated those gains so efficiently. Goldman according to their recent conference call, had (and is the main component of its defense) a Supersenior tranche on its books that had to be written down. However, this a relatively safe investments both in terms of subordination and length of existence (shortest amortization period). This has led to comments that claim the position was just a way to insure their innocence in the matter.

CONCLUSION

Despite claims by members of the financial community that Goldman’s derivative trades were dangerous, they did nothing wrong either by offering their ABACUS products of their investment practices. In fact, the only unethical component was their procedure and that they according to SEC allegations intentionally misled their clients. In concept Paulson’s  investment position was no different than had you sold your house in 2007 at the peak, rented for 2 years, and bought back your own house for a massive discount.

Experts have claimed that the Goldman defense will reference other CDO transactions to claim that they in fact did not mislead clients, but were as in most transactions preserving the anonymity of the long and short parties. The next upcoming weeks will determine whether Goldman is in fact guilty of fraud and whether they or Paulson & Company are liable for investor losses. The overall lesson in my opinion is that it pays to know who and why is on the other side of any trade.

Link to Pitchbook

Link to SEC Complaint