Speculation and Gambling; Goldman and Abacus

When I began working on futures market issues in 1980, it was explained to me that there was an important difference between “gambling” and “speculation.”  Gambling involves the creation of risk when none previously existed.  Speculation involved the transference of risk which already existed.  I thought about this and could see that a wheat farmer is long wheat and thus bears the price risk until harvest.  He could transfer that risk to a speculator by using the futures markets.  But, it seemed to me, that, when both the long and the short side are speculators, there is no risk transference.  The response to that is that the speculators not only serve to take on the risks of others but also provide liquidity to make this possible.  Obviously, while there is a legal difference, from the point of view of the speculator, there is little difference between what they do and gambling.  As a matter of public policy, we exempt some types of speculation from state gaming laws and provide them with more favorable tax treatment than that afforded to gambling.

This came to mind when I read accounts of the Senator Carl Levin’s subcommittee hearings on Goldman Sachs.  Senators compared Goldman to a casino; one even compared the firm unfavorably to the casinos in Las Vegas, because at least there the customers know that the odds favor the house.

At the center of this is a synthetic CDO that Goldman sponsored, ABACUS 2007-AC1.  The SEC has charged that Goldman and one employee, Fabrice Tourre, committed securities fraud by failing to tell the portfolio selection agent, ACA, and potential investors that John Paulson, who was involved in the selection process was planning to enter into a credit default swap by which he would benefit if the securities in the reference portfolio experienced credit events.

To my mind, the SEC complaint, if true, presents outrageous behavior by Goldman.  That is not to say, however, that the SEC can prevail in court.  For a discussion of the difficulties that the SEC may have if this case is not settled before going to trial, one can read Sebastian Mallaby’s column on this in the Washington Post and former SEC Commissioner Joseph Grundfest presentation at Stanford.  Nevertheless, what is legal is not necessarily right, and Goldman has already lost in the court of public opinion on this matter.  As smart and savvy as the people at Goldman are, I don’t think they can change the public’s judgment on this.

What about the gambling charge?  The offering circular and the “flipbook” for ABACUS 2007-AC1 are now publicly available on the web.  Without going into excruciating detail, the way ABACUS worked is that investors buy securities issued by a Cayman Island corporation created solely for this transaction.  The money raised is used by the corporation to buy a particular collateralized loan obligation (“CLO”) and in some cases other securities. The CLO is described on page 71 of the offering circular.  Its name is GWOLF 2007-1A A and its CUSIP number is 398078AB1. Reuters has an article indicating that Goldman may have used ABACUS to unload this complex security it had underwritten, which seems problematical.

The CLO serves as collateral for a credit default swap (“CDS”) transaction and provides income to the corporation which is part of the cash flow passed on to investors.  The corporation also enters into other derivatives transactions, including a swap to transform the income flow from the collateral security and a collateral put to be exercised in certain circumstances.

Th CDS references the portfolio of subprime residential mortgage backed securities.  ABACUS sells protection to a counterparty, in this case Goldman.  The protection buyer (Goldman) pays a premium at periodic intervals to ABACUS, which provides income to ABACUS to pay investors in its notes.

There are various classes of the notes.  If there are losses on the CDS, the lower classes take a hit before the higher ones.  The highest class is called “super senior” and the next class is called “A.”  Both these classes have a priority that the rating agencies said merited being rated AAA, on the theory that only a certain percentage of the underlying mortgages will default.  Goldman, as was the general practice, divided this higher tier into two.  The rating agencies do not give a rating higher than AAA, but the issuers of synthetic CDOs maintained that the super senior piece was virtually without risk.  Note that, because any losses that spilled into the category that the rating agencies rated AAA are first borne by the non-super senior piece, this division of the AAA piece in two makes that portion riskier than if the AAA piece had not been split in two.  The reason to do this division was to provide the non-super senior AAA piece a higher coupon than it would have if this division had not been made.  But even though the rating agencies put a AAA rating on this class, the reason it could get a higher coupon was that it was riskier than the super senior piece which got a lower coupon.

What appears to be the case with respect to the CDS involved in this transaction, Goldman entered into offsetting transactions with John Paulson’s fund and other entities.  Darrell Duffie of Stanford has some charts that show this.

The reason that synthetic CDOs such as ABACUS were created is that, in a low-interest rate environment, the demand for relatively high income from a supposedly AAA investment could not be met with conventional CDS structures.  Even with the amount of subprime mortgage loans that were being made, there were not enough to meet the demand.  In fact, it was useful for the investment banks that there were parties such as John Paulson around, who wanted to short this market because they thought there would be a real estate market collapse.  This enabled synthetic CDOs to be created.

One paragraph in the offering circular that should have warned potential investors of the risks they would be taking and the information asymmetry involved caught may attention:

“The Protection Buyer or its affiliates and/or the Portfolio Selection Agent or its affiliates may have information, including material, non-public information, regarding the Reference Obligations and the Reference Entities.  Neither the Protection Buyer nor the Portfolio Selection Agent will provide the Issuer, the Trustee, the Issuing and Paying Agent, any Noteholder or any other Person with any such non-public information.  In addition, neither the Protection Buyer nor the Portfolio Selection Agent will provide the Issuer, the Trustee, the Issuing and Paying Agent, any Holder of any Note or any other Person with any such information that is public (including financial information or notices), except in the case of information pertaining to one or more Credit Events with respect to each Reference Entity and one or more Reference Obligation(s) of such Reference Entity in connection with which the Protection Buyer is seeking payment of one or more Cash Settlement Amounts.” (p. 25 of the offering circular.)

At least in retrospect, this was more than mere boilerplate and should have warned investors off.  Goldman was saying that the odds may not be in the investors’ favor.

Except for the possible exception of the investment in the CLO, this transaction does look like gambling, since neither the investors nor the ultimate “protection buyers” in the CDS transaction likely had an existing risk they needed to hedge.  One side was betting that it could get high income because they did not think there would be a real estate crash serious enough to affect their investment; the other side was betting on a severe crash.

Whether or not this type of transaction should be permitted and given preferential treatment under federal law to casino gambling is a judgement call.  Libertarians and free market advocates would say that these transactions should be permitted; people should be free to enter into what transactions they want.  As to the allegation that Goldman was selling bad securities, they would say that even risky securities are worthwhile at a market determined price.  That some people were concerned about the real estate market was widely known, and the facts about that market were generally available.

On the other hand, one can make the argument that unlike gambling in Vegas, where what happens stays in Vegas (more or less), the wrong bets in the synthetic CDS market resulted in collateral damage affecting us all.  If that is not a convincing argument for prohibiting synthetic CDS structures, it is certainly an argument for better regulation.

At the moment, it may not matter much because I suspect the appetite for and issuance of synthetic CDS has gone way down.  But no one knows what will happen over the next decades.  The history of financial markets is that they do periodically get carried away and something nerve wracking happens, usually every few years.  Memories are short and new people are constantly entering into these markets.  It is not possible or even desirable to try to prevent all excesses, because one would be also prohibiting useful activity, but better regulation and tools are needed to make severe crises less likely and to handle those which do occur.