Submitted by David Fiderer
Dirty Little Secrets About Goldman's Collateral Calls on AIG
When it comes to AIG's liquidity crisis, Wall Street's conventional wisdom absolves Goldman from blame. Goldman's people, so the story goes, were smart and therefore prescient about the declining values of CDOs. So their demands for cash margin from AIG, which insured billions of toxic CDOs for Goldman's benefit, were legitimate. By contrast, AIG's people, the poster boys for financial incompetence, kept flailing about because they were in denial until everything reached a crisis point in September 2008.
Yes Goldman was smart, and yes, the people at AIG were clueless, which is why Goldman could pull off such an audacious scam. Goldman's demands for margin were made in bad faith, and possibly under fraudulent pretenses. The conventional wisdom overlooks a critical point: The legal documents had no teeth and might have been impossible to enforce.
The problems with the documents, in the context of the overall business deal, require a bit of explanation. But it's worthwhile to remember that all these deals are governed by two truisms: First, if you skip a step in analyzing a structured deal, you probably end up with the wrong answer. And second, almost everything about CDOs is kept secret in order to protect the guilty.
Transactions Designed to Prevent Any Sale On The Open Market
For a sense of how it all worked, consider West Coast Funding I, a $2.7 billion CDO that closed on July 26, 2006. This deal, which was underwritten and structured by Goldman, was exactly like 100 others taken on by AIG. West Coast I had seven rated tranches, of which the most senior four, representing 94% of the total deal, were rated triple-A. The two most senior tranches, A-1a and A-1b, were called "super-senior" because they ranked higher than the other triple-A slices. The exact amount of tranche A-1a was $1,187,950,000. The exact amount of tranche A-1b was $1,187,950,000.
One week before West Coast Funding I closed, on July 19, 2006, AIG
Financial Products entered into two credit default swaps with Goldman
Sachs International. The trades would be effective on July 26, 2006,
the date on which West Coast I first came into existence. Effective
that date, AIG insured $1,187,950,000, or 100%, of tranche A-1a, and
$1,187,850,000, or 99.992%, of tranche A-1b. The entire super-senior
tranche, about 88% of the CDO, was insured for the benefit of one bank
on the same day that the CDO closed or came in to existence. This was
the fixed template, from which there was almost no deviation, on
virtually all of the "multi-sector" CDOs insured by AIGFP.
There was never a single day when anyone, other than AIG, ever took
the direct credit risk on the super-senior tranche. The essence of the
transaction, the business model for all these deals, was to preempt the
possibility that the super-senior tranches could ever possibly be sold
on the open market. The West Coast I Offering Circular,
issued weeks later on September 16, 2006, gives no indication that 88%
of the credit risk was insured by AIG, which also extended
interest-rate swaps to the CDO. Again, almost everything is kept secret
in order to protect the guilty.
Goldman claimed that it "sold" these tranches to third parties, but
that half-truth is highly deceptive. After the deal closed, Goldman
held two assets: The super-senior tranches and the credit default
swaps. As the insured beneficiary, Goldman would never sell the tranche
free and clear on the open market. These particular credit default
swaps required physical settlement; Goldman was required to
physically deliver title to the tranche securities before it could
recover any insurance proceeds in case West Coast I defaulted. If
Goldman lost control of the super-senior tranches of West Coast I, its
swaps could become worthless.
So Goldman never sold West Coast I, or any of the other insured
tranches, free and clear to anyone else. The asset "sales" were
structured so that title would revert back to Goldman when it needed to
collect on the swaps. In the case of West Coast I, it entered into a
series of tie-in sales, package deals that assured Goldman's ability to
take back title to the asset in case an insurance claim ever came due.
Nominally it "sold" pieces of West Coast I to some obscure European
institutions, and simultaneously insured those pieces with credit
default swaps. The "buyers," ZurcherKantonalbank, and Depfa Bank and
Zulma (see Tab 39 of FCIC exhibits)
had no presence in the U.S. residential mortgage market. But they
weren't concerned. From their credit perspective, the package deal was
pretty much the same thing as buying a bond from Goldman Sachs, with
some extra collateral attached. At the point when the CDO defaulted,
the European "purchasers" would be reimbursed by Goldman, after they
physically delivered back to the investment bank.
West Coast I was unlike many other Goldman CDOs, in that the insured
beneficiary was also the bank that structured and underwrote the deal.
More often than not, Goldman deals were not insured for the benefit of
Goldman, but for the benefit of Societe Generale, which worked closely
with Goldman to create the illusion of a real CDO market, one where
banks went through the motions of buying and selling the super-senior
tranches. Goldman would sell the entire super-senior tranche on a CDO's
closing date, and SG would simultaneously close on an AIG swap insuring
the entire tranche amount. SG never really "bought" anything free and
clear, in that it never assumed one dollar's worth of direct credit
risk on Goldman deals for even one single day. Goldman "sold" about $11
billion worth of CDOs for SG this way. It also "sold" another $4
billion worth of CDOs to another French bank, Calyon. But the only
party that took direct credit risk on the CDOs was AIG.
Most of the deals insured for Goldman's benefit were underwritten
and structured by Merrill Lynch, which was largest CDO underwriter on
Wall Street. But again, Goldman "purchased" the entire super-senior
tranche on the same day that the CDO was created, and on the same day
that AIG insured the entire amount. It didn't matter how large or how
small the tranche was, Goldman bought the entire amount and AIG insured
the entire amount, and the swap required Goldman to hold on to the
tranche in order to collect insurance proceeds. So there was never a
moment, when any of these deals could ever be bought or purchased, free
and clear, on the open market.
The Impossibility of Measuring An Undefined "Market Value" When Neither the Assets, Nor the Market, Ever Existed
In making its demands for margin, Goldman calculated the value of
West Coast I as an asset that was entirely separate from its AIG
guarantee. But that asset had never existed in the real world. No one
had ever bought or sold those tranches severed from the umbilical cord
of a double-A corporate guarantee. When people talk about "market
value," they talk about an external reality check from a bona fide
market, one dominated by arm's-length purchase/sale transactions, and
one with a cash basis, where price discovery is ascertained by people
laying out real money. The CDO market, at least the market for the $70
billion worth of super-senior CDO tranches insured by AIG, was nothing
more than an illusion.
Was the concept of market value even relevant? Not according to
PriceWaterhouse Coopers, the auditor for both AIG and Goldman. It had
determined that, under current accounting standards, the fair value of
these CDOs could not be measured according to comparable sales or
market benchmarks. These were "Level 3" assets, meaning that the only
appropriate way a company could measure the assets' value was on a
fundamental basis, through financial models developed internally by AIG
Aside from the absence of any real market, or any commonly accepted
measure of market value, there was no definition of "market value" in
the contracts between AIG and Goldman. The term "market value," is
rarely self evident in swap contracts that reference financial
instruments traded in real markets. Those contracts don't simply
reference the market price of oil; they reference the closing price of
West Texas Intermediate quoted at close of the trading day on the
NYMEX. They don't simply reference Libor, but JPMorgan Chase 30-day
Libor. But the contracts between Goldman and AIG offered no guidance
for measuring the market value of the CDOs.
But even if we assume that someone could establish a theoretical
market value based on a company's internal financial model, which model
would prevail? Neither one. No single party decided what the market
value actually was; Goldman and AIG both decided. Both companies were
designated as the "Calculation Agent" in the swap confirmations. (AIG
was the exclusive Calculation Agent for swaps conducted with SG.)
Goldman and AIG could debate the value of West Coast I, and all the
other CDOs, and then, after failing to agree, each side would appoint
its independent representative, who would then take a while to
deliberate on the valuation. If they failed to agree, the two
representatives would deliberate over the choice of a final, single
arbiter, who would then ascertain the value of the CDO tranche, at a
point long past the date of the initial dispute.
It wasn't only the "market values" that might shift over time; the
investments within the CDO portfolios could easily shift as well. Most
of the CDOs insured by AIG were managed, meaning that they were run by
asset managers who had great latitude to substitute different
investments within the CDOs. So each time the assets changed, a
brand-new valuation was necessary. So nothing about the CDOs'
valuations, in a nonexistent market, could ever be pinned down. So
nothing could ever be settled, except in court or in arbitration.
And if there's no way to attain a timely resolution, then the margin
requirements seem meaningless. In real markets that trade with a cash
basis, the need for timeliness can be easily measured. If you pay less
attention to the form and focus the underlying substance of the
transactions, which were governed by vague and sloppy wording, you have
to wonder whether Goldman's demands for margin would have been
But Goldman's management didn't care about the contract language. It
wanted AIG to hand over cash, sooner rather than later. And they pulled
out the stops to pressure into everyone into accommodating their
How Goldman Tried to Scam AIG
A few days after executives at Goldman made their initial demands in
July 2007, they started playing hardball with Tom Athan of AIGFP. "Tough conf call with Goldman," he emailed afterwards.
"They are not budging and are acting irrational." The issue was a "test
case" that had gone to "the highest levels" at Goldman. But--here's the
smoking gun--Goldman insisted that AIG present it "actionable firm bids
and firm offers" to come up with a "mid market quotation," for valuing
the insured assets. The idea was patently nonsensical. There never were
and there never would be any bona fide bids of any kind, be they
"actionable and firm," or merely "indicative," because no one ever
intended to buy or sell the assets, which were never available for sale
on a free-and-clear basis. But the people at AIG never put two and two
together. Like a bunch of suckers, they went out soliciting "quotes,"
which Merrill Lynch would only offer on a confidential basis. They also
got quotes from Goldman and SG.
Goldman clearly wanted to override the wording of the documents,
which prevented the bank from asserting the upper hand, according to
reporting in The New York Times:
[David] Viniar, Goldman's chief financial officer, advised
the insurer in the fall of 2007 that because the two companies shared
the same auditor, PricewaterhouseCoopers, A.I.G. should accept
Goldman's valuations, according to a person with knowledge of the
When A.I.G. asked Goldman to submit the dispute to a panel of
independent firms, Goldman resisted, internal e-mail messages show. In
a March 7, 2008, phone call, [AIGFP CEO Joe] Cassano discussed
surveying other dealers to gauge prices with Michael Sherwood,
Goldman's vice chairman. At that time, Goldman calculated that A.I.G.
owed it $4.6 billion, on top of the $2 billion already paid. A.I.G.
contended it only owed an additional $1.2 billion.
Mr. Sherwood said he did not want to ask other firms to value the
securities because "it would be 'embarrassing' if we brought the market
into our disagreement," according to an e-mail message from Mr. Cassano
that described the call.
The Goldman spokesman disputed this account, saying instead that
Goldman was willing to consult third parties but could not agree with
A.I.G. on the methodology
Nonetheless, it sure looks as if Goldman has been trying to scam the
Financial Crisis Inquiry Commission in much the same manner that it
scammed AIG, with a lot of dissembling about "market values." Again,
Goldman's auditor, PriceWaterhouse and, Deloitte & Touche,
the auditor for Maiden Lane III, the entity holding the toxic CDOs once
insured by AIG, found that these assets had no "market value;" they
could not neither be measured according to actual sales nor comparable
sales of similar assets in the marketplace. The only valid way to ascertain their fair value would be to perform fundamental cash flow analyses.
Why Goldman Never Shares Its Cash Flow Calculations With Anyone Else
There's no indication that Goldman shared its cash flow evaluations
with AIG or with anyone else. A lot of documents have been released so
far by Goldman, the FCIC, the Senate Permanent Subcommittee on Investigations, and the Congressional Oversight Panel.
None of those documents show any contemporaneous communication by Goldman to AIG
about how the bank actually calculated the values of the CDOs. In its
highly selective and misleading history of events written
after the fact, Goldman glosses over the fact that the reference assets
were never available for sale on the open market, and that its
valuations based on comparable sales violated fair value accounting
There's no doubt that Goldman performed fundamental cash flow
analyses on all its mortgage assets. As management relayed to the Board
of Directors in a slide titled, "Independent Price Verification" :
A dedicated group within Controllers performs an
independent price verification of the mortgage inventory. The team is
highly specialized and has extensive experience in the valuation of
mortgage related products...Price verification analysis utilizes four
core strategies, [including] Fundamental analysis [which] utilizes
discounted cash flow (DCF), option adjusted spread (OAS) or
securitization analysis. Observable market data or inputs are
incorporated when available and appropriate.
Almost certainly, the disclosure of those cash flows would have been
damning, not because Goldman had overstated the declines in values of
the CDOs insured by AIG, quite the opposite. It might have revealed how
Goldman and others had been artificially propping up the values of
toxic securities, as part of an elaborate pump and dump scheme.
Two years ago, Sylvain Raynes, formerly of Goldman and Moody's and currently a principal at R&R Consulting,
replicated the type of cash flow analysis done by Goldman and every
major Wall Street bank. By statistically evaluating the cash flows of
the 20 subprime bonds referenced in the ABX 2006-1 indice s, Raynes calculated their composite values beginning in January 2006, when the ABX was first launched. Beginning in January 2006,
all of the indices, aside from the AAA index, were worth pennies on the
dollar. In December 2006, Goldman closed on its new $2 billion CDO, Hudson Mezzanine Funding I,
which insured all of the BBB and BBB- bonds in the ABX at 100 cents on
the dollar. In the months that followed, Goldman offered up ABACUS 2007
AC-1, Anderson Mezzanine Funding, and other toxic CDOs that were designed to fail.