A knife fight is not a mediation

The excellent Ezra Klein has ruined his electoral career for nothing. That’s a shame; I’d vote for him. Ezra writes:

If an investment bank is structuring a trade for two clients, it has an obligation to serve its clients. That is to say, it needs to structure the trade they want to be part of and disclose all relevant information necessary for them to evaluate the trade. But if the firm, or the employees structuring this trade, think that one side is going to win and the other is going to lose, I don’t think they have an obligation to warn the losers… The SEC’s case against Goldman simply says that they failed to disclose relevant information that one side needed to decide for themselves whether going long on the Abacus deal was a good or bad trade. That is to say, the issue isn’t whether Goldman acted in the client’s best interest but whether they made it unnecessarily difficult for the client to act in his own best interest.

Goldman wasn’t structuring a trade between two clients, as far as IKB and ACA were concerned. It was working to form a business entity called ABACUS 2007-AC1, LTD and underwriting an issue of securities by that entity. The only clients formally involved were IKB and ACA, and they were on the same side of the deal.

If this had been an adversarial deal, Goldman would have had no obligation to inform the side that wasn’t paying it whether they were making a good trade. But if this had been an adversarial deal, Goldman would have been advising one party or the other. Both parties could not have been its customers.

Imagine you are trying to buy a house. It is contentious. Disputes arise over price, warranties, settlement terms, etc. You would hire an agent, and the other party would hire an agent. Those agents would be different people. The hazards of relying on the same advisor in a difficult negotiation are obvious.

IKB/ACA may have been “sophisticated”. They may have been dumb, or corrupt, or unlucky. But, in an adversarial negotiation with John Paulson, they would not have shared the same agent with him. A knife fight is not a mediation.

The whole issue is that IKB/ACA did not know that they were in an adversarial negotiation and that the other guy had Goldman Sachs as its agent. They thought Goldman Sachs was working for them, underwriting securities of a special purpose entity it was putting together to satisfy investor interest. If IKB/ACA had been negotiating a very complex $192M custom trade with John Paulson, there would not have been a “flipbook” and a “prospectus”, just sign the dotted line. There would have been conference rooms and long hours and thousand-page paranoid contracts scrutinized and initialed in triplicate.

There is no circumstance where an investment bank “structures a trade for two clients” whose interests are opposed when the terms are anything but standard. I mean, really. Think about it. It’s Orwellian — Goldman calls a practice that is absurd on face “market making”, and suddenly it’s normal except for technical questions about who picked what securities or who should have suspected something.

What happened here is nothing like what a market maker does. A market maker takes the other side of client-initiated trades, and then lays off the risk. ABACUS was initiated and sold by Goldman Sachs, at a hidden party’s request. Goldman was unwilling to make a market for Paulson at a price he would have accepted, so it manufactured an entity willing to do so. Investors in that entity were not informed that they were dealing with an active, involved adversary. And Goldman has the nerve to call both sides of the arrangement “customers”.

This was a high finance version of the same pump-and-dump schemes you get by e-mail. Paulson needed buyers for what he was selling. Goldman sent around the flipbook until it found some, and without revealing that a hidden counterparty wanted to dump. This is not an ethical practice. I don’t know whether it’s illegal. But if really smart, well-intentioned people like Ezra can’t quite see that it’s disreputable, if it seems like a he-said, she-said technical kind of thing, we are in deep trouble.

P.S. For this deal to be okay, Paulson’s role would have had to be disclosed plainly and in writing. His name need not have been mentioned, although it would not have remained hidden for long. But Goldman would have had to reveal that a party wishing to take a large short position had initiated the deal and would be involved in its design. Goldman would also have had to make clear, in writing, that this party was its client.

Like many Goldman apologists, I suspect that by the time the deal closed, some of the ACA guys probably knew or had guessed what was going on. Maybe the IKB guys knew too. That doesn’t matter. Individuals taking bonuses for deals at ACA and IKB were not the “investors”. ACA and IKB’s shareholders were the investors, and ultimately British and German taxpayers. Goldman had an obligation to put important facts in writing. By not doing so, Goldman created plausible deniability for employees at ACA and IKB who had a personal interest in closing the deal. The wink-wink/nudge-nudge act mitigated career risk, helping to enable corrupt stupidity. Informal disclosure does an end run around risk managers at both firms, who would have expected discussion of an active, adversarial counterparty in the documents they reviewed. Even if some people at ACA knew, the deal might never have gotten done had ACA or IKB formally known. $192M deals that become $1B deals should be fully documented, in ink.

Update History:

  • 28-April-2010, 10:10 a.m. EDT: Changed “sharing the same agent” to “relying on the same advisor”.
  • 29-April-2010, 7:00 a.m. EDT: Shortened first link to Ezra Klein’s piece. Removed one “really” where there were two in last sentence before postscript.

99 Responses to “A knife fight is not a mediation”

  1. Sandrew writes:

    “The whole issue is that IKB/ACA did not know that they were in an adversarial negotiation and that the other guy had Goldman Sachs as its agent.”

    “I suspect that by the time the deal closed, some of the ACA guys probably knew or had guessed what was going on. Maybe the IKB guys knew too. That doesn’t matter.

    Seems like a contradiction. Is the distinction in your mind one of timing?

    Does the ethical case hinge on whether Goldman was forthright with ACA and IKB wrt the adversarial nature of Paulson’s prospective interest? (I know the legal case certainly does, but legal and ethical standards don’t perfectly overlap.) If so, does it matter whether Goldman shared this information with designated agents of ACA/IKB as opposed to the principals thereof? Does it matter the medium in which they communicated this information (specifically as a matter of ethics, not merely prudence)?

  2. Sandrew writes:

    “There is no circumstance where an investment bank ‘structures a trade for two clients’ whose interests are opposed when the terms are anything but standard.”

    “What happened here is nothing like what a market maker does. A market maker takes the other side of client-initiated trades, and then lays off the risk.”

    I agree with your description of market-making, a description this deal does not fit. I would characterize Goldman’s role here as a broker, which is perhaps too fine a distinction to matter. I think your assertion is that Goldman was holding itself out as an agent to both parties, when in reality it could only be agent to one (Paulson). I’m not yet convinced this is the case.

  3. Danny Black writes:

    By all its own account, Paulson’s hedge fund TOLD ACA that it was short and one only had to read the news at the time to see Paulson was proclaiming their view. I think when full disclosure comes out that it will turn out that ACA thought it was the one taking Paulson for a ride.

    I wonder if the SEC is going to sue all the people who structured a short bet that **lost** money instead of post-hoc being “the greatest trade ever”.

  4. Kid Dynamite writes:

    i thought the most interesting thing to come out of the hearings was Blankfein’s clear statement that their obligation is to make sure the products that they sell their clients do exactly what the products are supposed to do – and I agree with Blankfein on this. This is very different from making sure that the products provide returns for their clients, or from offering their clients views on how GS thought the securities might perform. Of course, the Senators don’t understand this, and don’t understand that fiduciary duty isn’t relevant here. I’m guessing that’s what Ezra meant when he said “I don’t think they have an obligation to warn the losers… ”

    of course, it was also interesting that none of the Senators brought up the point you mentioned above about how CDO origination is quite a bit different from the kind of market making to which Blankfein repeatedly referred.

  5. […] Steve Randy Waldman, “A knife fight is not a mediation.”  (Interfluidity) […]

  6. Jon W writes:

    Probably where many of the differences of opinion form is whether or not we think of offerings as market making activities or as some kind of “sale/promotional.” I think the diffculty in this discussion is that for most people when they see a pitchbook or a sale of offerings, they think of it coming from an advisory standpoint. People find the behavior appalling because the assumption is that the sales function of goldman and other banks is to present you with things that will be “good” for you and make you risk adjusted money. That is a perfectly reasonable view to have and I would be appalled too if I had that view.

    From my perspective and the perspective of GS, their function is not to make you money. At least not in respect to trading, market making and even offerings. The presentations, pitchbooks etc simply exist to tell you “this is what you will be buying, and this is a price at which you can buy it for.” In theory the IPO prices in an offering are supposed to be values that reflect a neutral point between the sellers and buyers. Even here there is an adversarial relationship. Sellers want more, and buyers want less.

    As much as GOLDMAN doesn’t want to admit it, most of their operations are as bookies not as advisors. And being a bookie isnt a bad thing. Lots of professions do basically the same thing. A real estate broker is no different. They’re just guys that know the process , collect a fee and and set some lines.

    It’s why you heard GS execs saying they sold exposure, not returns.

    Even in an origination the disclore and legal / ethical standards arent to tell investors they will make money on the offering. They’re there as sa stamp of approval saying that the issuer, in this case to make sure the founders/shareholders of the company aren’t misrepresenting the facts on the underlying assets / business. Here there is a ton of material, non-public information that the original shareholders have possession of and hence the job othe investment bank.

    In an ABACUS transaction no such thing exists regarding the underlyers. What Paulson has is an opinion, not control of inside information regarding the assets. The prospectus outlines what the risks are, what the cashflows are, and how these change with respect to your returns. Unless Paulson has material non public information regarding the performance of the underlying securities, how can his opinion and intentions be material to the offering?

  7. csissoko writes:

    SRW: I really think you’ve succeeded in identifying the real problem here.

    If Goldman had told IKB: We are trying to lay off the risk of a bespoke CDO, are you interested? there would be nothing to complain about. [Note that there was no need to mention Paulson’s name.]

    To the degree that structured finance products are created for the purpose of hiding what the market maker is actually doing, they are fraudulent — even if they happen to be legalized fraud.

  8. cudBwrong writes:

    The basis of the SEC lawsuit, according to their press release, is not just that GS omitted a disclosure, but that they made explicit statements which were contrary to fact.

    The SEC press release quotes SEC Enforcement Division Director Robert Khuzami as follows:

    “Goldman wrongly permitted a client that was betting against the mortgage market to heavily influence which mortgage securities to include in an investment portfolio, while telling other investors that the securities were selected by an independent, objective third party.”

    The important part of this quote is the final clause, the claim that Goldman was telling other investors something different.

    This is supported by the statement in the flip book:

    “The Portfolio Selection Agent has selected a target granular Reference Portfolio”

    when it is alleged that the portfolio components actually were selected, in whole or in part, by Paulson.

    On this basis, it’s possible to argue the very descriptions of ACA as the “Portfolio Selection Agent,” and of their compensation as a “Portfolio Selection Fee” were misrepresentations.

    What is securities fraud? It is the attempt to induce an investor to make decisions on the basis of false information.

  9. anon writes:

    “The whole issue is that IKB/ACA did not know that they were in an adversarial negotiation and that the other guy had Goldman Sachs as its agent.”

    At which point it became ACA’s responsibility to figure out why ACA and Paulson were actually at the same table agreeing on the collateral selection. What did they think the reason was? And where did they think Paulson came from if not via Goldman? Duh…

  10. Robert writes:

    Thanks Steve, your analysis is excellent, I enjoy reading it.

  11. JKH writes:


    I’m late to this, but several thoughts:

    To the extent there is an adversarial relationship between the short and the long, it seems to me this relationship is entirely manifested in the portfolio selection process. ACA controlled the portfolio selection process. Both the longs and the shorts were represented in the portfolio selection process. I think the interesting paradox is that in theory its not necessary for those involved in the portfolio selection process to know each other’s roles, provided that those adversarial pricing interests be represented. The further paradox is that it is actually constructive that these adversarial interests be represented in that process for the sake of fair risk construction and price discovery. The value added is that the interests of short and long positions be considered in constructing and pricing the deal, which can be achieved by such a process – not that the actual roles be disclosed, although such a set of circumstances is bizarrre. That said, I believe the case will demonstrate that these adversarial interests were either disclosed or should have been evident by an intelligent particpant in the dynamics of the process. In any event, ACA and not Goldman had the responsibility for ensuring a fair portfolio selection process once it had been designated as the CDO manager.

    Which brings me to the point of Goldman’s role as agent in this case. I think you are mistaken in identifying the character of the agency role here. To the degree that portfolio selection was not Goldman’s responsibility, it could not have been Goldman’s role to represent either shorts or longs directly as agent in the fairness of risk selection and pricing. That was ACA’s role. It was Goldman’s role to bring parties together and to ensure a fair process for the construction of the deal. Again, that actually didn’t formally required the disclosure of adversarial roles – only the representation of adversarial roles in portfolio construction and pricing. Beyond that, economic fairness was ACA’s responsbility. Therefore, I see no conflict in Goldman’s role as dual agent for both the shorts and the longs in terms of bringing these counterparties together in order to structure a deal that was economically transparent by selection of the underlyings and pricing of the deal. And despite the awkwardness of the issue around counterparty role identification, I believe they did that via ACA’s portfolio selection control function, which engaged the economic interests of both the long and short side, notwithstanding this other counterparty noise problem.

    Finally, I don’t think that the debate over a market making function is central to the ABACUS discussion. It was certainly central at times in yesterday’s hearings. The proof that it was central was the tracking of Goldman’s risk position as a firm and their explanation of how their participation in the CDO market – particularly taking the short side of newly issued CDOs during 2007 – was used to reduce long risk and go short. The proof of a market making function is in the evidence of a changing net risk position. Perhaps I’ve missed some important information here, but it’s not clear to me that Goldman has pressed the case regarding a market making rationale for the ABACUS transaction in question. That said, the small equity piece does qualify as a market making component (and I thought there was a timing issue on the ACA super senior where Goldman carried it for a while, but maybe I’m mistaken on that).

    I agree with comments by Kid D. and Jon W.

  12. Let me spit this out in my usual Village Idiot way. A few commenters have pointed out that GS testifiers were constrained by the advice of attorneys. I agree, but still find it useful to notice their general line of speech.

    From where I sit, GS is in a bind. They can certainly claim that they met the standard of Full Disclosure, but people like me won’t buy it. For one thing, I learned about Full Disclosure when I sold my house, and had to write a list of every little problem that I was aware of. They seem to have left out facts that might well have influenced the deal. They can say that they met the requirements, but then it just looks like they structured the deal in order to avoid full disclosure. That might be legal, but it puts a huge Buyer Beware sign on them.

    And that brings me to Caveat Emptor. Throughout this Crisis, various people have been trying to bring back Caveat Emptor, as opposed to Full Disclosure. The clearest proof of this is when they start labeling the sucker as “sophisticated”, “not mom and pop”, etc. The only reason to resort to this line of defense is because you’re worried about whether or not you met the standard of Full Disclosure. By crying Caveat Emptor, you’re attempting to rhetorically lower the standard.

    The difference between now and 1933, when the Securities Act was passed, so far as I can determine, is that people back then mocked this line of rhetoric, but, today, many people buy into it. There’s probably been a shift in mores over the years on this, also evidenced in the fact that, in 1933, some of the most important Free Market Advocates favored, among other things, 100% Reserves, which they considered a very strict system, while many of today’s Free Market advocates call that view, as Peston said, Nutters.

  13. Distant Observer writes:

    It seems perfectly obvious what happened:This deal had a story and a history. GS omitted that story and history from the offering documents for a very simple reason: If it had told the story no one would have bought into the deal. Too irregular;too dicey; if it goes bad the Board or the lawyers or the compliance department will fry me. That is why GS obfuscated from the beginning, and why now, its defense tries to move the goalposts by claiming it was acting as a broker,and not an underwriter, where, as earlier comments have noted, the standards of affirmative disclosure are much higher.

  14. carping demon writes:

    Caveat Emptor is a warning. It is not an imperative. It is not a standard, a law or a rule. It is not a philosophy, nor is it a POV. It is neither an argument nor a defense. It is a warning. It is as the southern end of a northbound diamondback, and in no way alleviates, remedies, justifies, ameliorates, reduces or excuses the poison at the northern end. No matter what the snakes say.

  15. Morgan_03 writes:

    Steve, which parties have referred to ABACUS as bespoke? I understand that Paulson referred to ABACUS as an “off the shelf” product, and that “his” deal was the 25th iteration of ABACUS. Maybe just semantics. As always, nice work.

  16. […] <!–Considerably more informed opinions on the Goldman ABACUS deal have been written on the interfluidity blog; I recommend the posts of April 25,  April 27 and April 28.–> […]

  17. […] – Abacus 2007-AC1 as pump and dump. […]

  18. Steve Randy Waldman writes:

    Sandrew — Re disclosure, the issue is what constitutes clear, meaningful disclosure to a corporate entity. (The issue isn’t timing, as long as the clear disclosure would have occurred prior to a commitment by investors.)

    As a reducto ad absurdam, suppose that the fab Fab, walking through the spotless halls of ACA, happened upon the janitor and struck up a conversation. “Yes,” he says to Bob, “we are doing this deal where a hedge fund guy — see that guy over there, he’s Paulson — will short credit, and some investors, perhaps even ACA will go long.”

    “Hmm”, says Bob, as he picks up his mop. “That sounds great.”

    Would that have constituted meaningful disclosure to ACA? Now for some purposes, Bob may be a designated agent of the firm. Bob is permitted to place and sign for orders at the janitorial supply company. But Bob is not generally involved in, for example, risk management.

    Obviously, I’m overstating. If Fab or Pellegrini had told the janitor and that was the basis for their claimed verbal disclosures, any jury would consider their literally true statements worse than lies. Let’s suppose they are not that slimy, and they did tell the people they were going over portfolio selection with.

    How is that different than telling the janitor? Because credit analysts are better paid, white-collar, “professionals”? The credit analysts may have some input into whether the deal gets done, but they are unlikely to have been the responsible decision-makers. Disclosures delivered verbally to arbitrary individuals at a firm can not be presumed to be reliably delivered to the “corporation”. That’s why there are very clear norms of intercorporate communication. Most human communication begins informally. But once understandings are reached on consequential matters, intercorporate communication is crystallized into writing, and efforts are made to ensure distribution to appropriately designated decision-makers. Written documents have several important properties: they serves not only to communicate, but to document communication. They can be delivered to multiple parties, and if delivered to an inappropriate person, they create a duty to forward that is risky to ignore. It is one thing for an individual to “forget” to inform some person in another department of some detail they heard in a conversation. It is quite another to receive a letter or a memo that contains information required by risk management and to then fail to forward it on. If things go wrong, the sender produces the document and knows to whom she sent it.

    Every corporation has both informational sprawl and conflicts of interest. Of necessity, legal and corporate norms ignore these for some purposes. A corporation may be responsible for commitments made by an employee, even if that employee was not authorized to make the commitment, if the person to whom the commitment was made had reasonable grounds to believe that the employee was so authorized. But “reasonable grounds” is a scale that slides with the scale of the commitment. When Bob orders janitorial supplies, the firm has to pay for them even if it turns out that Bob wasn’t authorized to, um, do those deals. But if Bob agrees on a handshake or even signs for a $42M CDO deal, that won’t cut it, the firm won’t be bound. Matters of great importance are performed in writing, and it is the duty of both parties to make some reasonable effort to ensure that appropriate agents consent to them. There are limits to this duty — it would not have been Goldman’s job to verify that ACA’s internal procedures for risk management were followed. But it would have been Goldman’s job to ensure that the person consenting was not a janitor or junior credit analyst. Similarly, Goldman had some duty to make disclosures of material information in a manner that could reasonably be expected to reach responsible parties. It had a duty of prudence to its shareholders to make material disclosures in a manner that could be evidenced and verified. This is very rarely an issue, because disclosures of material information are always made in writing.

    These issues are particularly important at financial firms, where conflicts of interest are acute. Compensation is often linked to doing deals, and risk managers can only say no. It is easy for a trader or portfolio selector to be overconfident, to persuade himself that the deal will be great and what the bean-counters don’t know won’t hurt anyone. Managing those conflicts is a firm’s own responsibility. Goldman cannot be held responsible if individuals at ACA evaded internal procedures. But Goldman did have a responsibility to conduct it’s business according to professional norms by disclosing material information in writing. No procedures would be sufficient at ACA if Goldman revealed potentially adverse information only in informal conversation and to individuals the greatest incentive to overlook any problems.

    Here’s a specific wrinkle that illustrates some of these points. The people Fab and Paulson had meetings with were involved in the portfolio selection process. They were from ACA Management. The firm ACA Management would earn a fee regardless of how the portfolio performed. ACA Management’s incentive was to get a deal done. ACA Capital, a different though related firm ended up investing $42M and insuring $909M of the deal. Was it reasonable for Goldman to expect that an informal discussion with employees at ACA Management constituted adequate disclosure to ACA Capital?

    Even if “ACA” had been a single legal entity, undocumented verbal disclosure would not have sufficed. But the specifics here remind us that, though not all firms are vampires, all are squids. They have many tentacles. Meaningful disclosure has to be made in a manner that has a reasonable shot of finding its way to the head, and that can be verified after the fact.


  19. Steve Randy Waldman writes:

    Note: This response owes something to the very excellent comments of nadezhda at The Baseline Scenario.

    Sandrew2 — I’m sorry you’re not yet convinced, but you have the structure of the argument. A broker is an agent of a seller or a buyer. A market-maker trades on its own account. An underwriter intermediates between a buyer and a seller in a manner that is inherently conflicted, but which compels very detailed disclosure particularly of the identity of the seller and its economic interest both before and after the deal.

    Goldman claims to have been a market maker, because that is the only role in which a certain kind of anonymity is the norm, and even there the anonymity is a bit of a chimera because the counterparty is the investment bank itself, trading on its own account. But a market maker need not reveal to whom or even if it eventually lays off its position.

    However, Goldman was not a market maker, because it played a large part in initiating the trade and did so on behalf of a third party. It was a broker, a seller’s agent. Goldman was also an underwriter, because it underwrote a new securities issue. There are no norms of anonymity in either of those two roles, and in underwriting, because there is an inherent conflict of interest, the norm is that not only are the identities of anyone who might be construed as a seller revealed, but their stake in the deal both before an after are crucial to the process. When IPOs are underwritten, investors want to know that they are not having a lemon dumped upon them. Usually they do so by insisting that the current owners retain substantial ownership in the firm after the deal is done, and by forcing current owners to agree to “lock-in period” during which they cannot sell. The conflict of interest, then, becomes fairly manageable, because after the sale, the interest of old and new investors are aligned: they all want the firm to succeed. [Prior to the IPO, there is a theoretically zero-sum struggle over price, which old investors almost always lose. There is a vast literature that tries to explain IPO underpricing, but to the degree that equity underwriting is rigged, it seems to be rigged in favor of new buyers, not old owners. However only new buyers who purchase the issue directly from the underwriter or issuer “win” on average. People who buy on the secondary market after the “IPO pop” usually overpay.]

    It is not uncommon for an investment bank to be both an agent of a seller and an underwriter. In fact it is the most common case — underwriters usually are sellers’ agents. But an underwriter/sellers’ agent has a duty to disclose information about the sellers.

    This deal appeared to be a less common case, where a firm was de novo and there was no seller. Yes, there were going to be sellers of credit protection, but that’s mostly meaningless. Suppose the de novo entity, instead of taking CDS positions, was going to buy stock like a closed-end fund? There would be “sellers” of the new entity’s portfolio, but there would be no active economic agent on the other side of the deal: stock would be purchased from numerous parties in public markets, with the investment bank acting as market maker. Similarly, ABACUS investors had no reason to think there was a meaningful seller in this deal.

    So, investors in cooperation with Goldman Sachs were going to incorporate a firm and sell it to themselves. In this kind of deal, the underwriter’s duty would be the investors (who are presumably paying the IB fees to structure the new entity). That’s what IKB and ACA would reasonably have assumed, absent disclosure of the actual history and nature of this deal. Who the sellers are, what there economic interests are, and that they exist is always material when securities are underwritten.

  20. Steve Randy Waldman writes:

    Danny — It doesn’t matter who ended up gaining or losing from the trade. Goldman’s behavior was unethical and unprofessional on the day the deal closed, before any credit events or price movements. Admittedly the SEC would probably not have pursued Goldman if the deal went the other way. It’s hard to win a legal case when there seem to be no damages, and there would be no political appetite. But the ethical questions I’m concerned with have nothing to do with how the investment ultimately performed.

  21. Steve Randy Waldman writes:

    KD — Repeating my comments to Danny, it doesn’t matter who won or lost. Goldman behaved unethically before the prices moved. Goldman did not have a fiduciary obligation to investors, but it had a clear obligation to disclose the existence and nature of another party intimately involved in the deal.

    I didn’t see the hearings. (Apparently my cable company has decided CSPAN3 is a subscription service, and the online feed kept blocking.) I expect that Senators said many silly things and got the accusations wrong, while the Goldmanites mischaracterized and misdirected. It might have been worth watching for the comedy, or perhaps my cable company saved me from a frustration-induced aneurysm.

  22. Steve Randy Waldman writes:

    Jon W — I agree with almost all of what you say, until the last two paragraphs.

    When IB’s offer pitchbooks and prospectuses and stuff, they are indeed acting as salesman, not advisors. Even when they are acting as advisors, no one can guarantee an investment will be good for you or make money. It is a risky world. When IBs act as advisors, they have a duty to give the best advice they can. If they take a fiduciary role (e.g. when they manage your money), they have an obligation to put your interest before their own. But when they are underwriters — salesman — their duty, as you say, is only to disclose, to accurately characterize what it is they are selling.

    But when selling a business, the economic interests of other stakeholders, both before and after the sale, are always relevant. Businesses are not sold by market-makers, they are sold by brokers and underwriters, and there is no presumption of anonymity because businesses would not sell on an anonymous basis. No matter how accurately an investment bank characterizes what the business hopes to do, what its inventory, assets, or personnel are, everything that can be known, the motivations and interests of the sellers are material. The world is inherently an uncertain world, and people (accurately) recognize that there is information content in others’ consequential choices. All buyers of new businesses demand that information, and often refuse to buy if the seller doesn’t keep “skin in the game” or offer warranties post-sale.

    Even for very standardized products like stock, counterparties require information about the motivation of the other side. There would be no market makers of common stock if it were impossible to distinguish between buyer- and seller-initiated trades and observe directional commitment as proxied by flow imbalance. (See my response to a previous comment of yours here if you haven’t, last paragraph.) Market makers are extremely keen to suss out the motivations and intentions of counterparties (who they view as a statistical swarm).

    The question of who is initiating a trade and who is supplying liquidity in response is always material. The liquidity supplier rationally takes into account the other party’s interest and builds in a margin of safety against adverse selection. This is true for the simplest of assets, let alone for a deal like this.

    Goldman masked the existence of an organized counterparty, its economic role in the transaction, and the fact that it initiated the trade. All of that information was material.

    So we agree on what Goldman’s duties were and were not in broad terms. But I’m certain they did not fulfill those duties, while I suspect that you may still disagree…

  23. Steve Randy Waldman writes:

    csissoko — Yup! If they had clearly stated that they had a client who wanted the short position and they were looking for someone to take the other side, that would have been fine!

  24. Steve Randy Waldman writes:

    cudBwrong — Yes. What the SEC alleges is stronger than what I’m describing. I hope it’s obvious that if and to the degree Goldman did actively lie, it behaved unethically (and probably illegally).

    But even Goldman never made a false statement, what it failed to disclose under the circumstances of this deal is enough to render its behavior sleazy.

  25. Steve Randy Waldman writes:

    anon — i have little sympathy for IKB / ACA (and a lot of sympathy for the German and British taxpayers who paid their tab).

    but it was Goldman’s duty to disclose. they were not required to ask.

    it is quite possible, as i described in the post, that employees at ACA (and possibly IKB) were corrupt and/or overconfident. if so, Goldman’s failure to disclose offered cover, plausible deniability, enabling a deal to get done which would surely have been canned by risk-averse bureaucrats if all the facts were on the table.

    the laws and norms that govern the Securities industry don’t exist just to protect snakes from one another, or innocent investors from snakes. they exist to also to help prevent corrupt collusion that victimizes those who aren’t party to their trades. there are lots of scenarios by which individuals can collude to extract rents from the rest of us in the financial system by looting the organizations they allegedly serve. only by forcing material information to be documented and creating reputational and sometimes legal jeopardy for those who violate their duty of loyalty to their employers can we prevent that misbehavior.

    the SEC isn’t suing on behalf of ACA or IKB. it is suing on behalf of the rest of us.

    i am pushing this case so hard not as a vendetta against Goldman or because ACA or IKB deserve some kind of “justice”. i wouldn’t care, if Goldman simply admitted that its practice was unethical, that it understood its duties going forward, and would no longer do this kind of business. that it refuses to do so, that it is fighting to legitimize harmful and unethical practices while claiming to serve clients, is unfathomable, and must be fought.

  26. Steve Randy Waldman writes:

    Robert — thanks for reading!

  27. Steve Randy Waldman writes:

    JKH — Both shorts and longs were represented in the portfolio selection process. But the longs did not know that shorts were represented. The marketing materials they received led them to believe quite the opposite.

    There are lots of ways to structure a market. One way is a typical anonymous public market with continuous limit order books. Another way is to get buyers and sellers to meet and negotiate price and terms. I suppose that another would be to put buyers and sellers together while disguising the role of the seller, but I’m skeptical of a theory suggesting that would lead to “fair risk construction and price discovery”.

    My view of this deal is that it was entirely about thwarting price discovery. That will probably be the subject of the next post, but you can usually anticipate my thinking. I suspect you understand the case. In any other market, the initiator of a trade inevitably pays a “lemons premium”, because the respondent, the party supplying liquidity that another party has demanded, has to wonder what the other party knows. That moves the market price. Of course, most traders, whether at Paulson’s level or a retail neophyte, know less than they think they know. But that leads to “noise”, back and forth trades about evenly balanced. Prices oscillate but do not shift. As consensus and conviction arise prices do move, frustrating the traders who think they know something (or who share a common misconception). No one likes the market to move against them before they can make their trades. But price shifts in response to unbalanced trade interest are precisely how markets are supposed to become informed, to get the prices of things approximately right. Traders who know (or think they know) something hate this, and the big clever ones try to thwart it. One technique is to persuade a firm you’ve invested in to repurchase shares while you intend to sell. Another is to talk down an asset that you intend to buy. And a third way is to talk up an asset while you intend to sell. That’s effectively what Paulson and Goldman connived to do. They did not passively set up a neutral structure, and wait for both longs and shorts to come and make suggestions about the portfolio. (How would those be adjudicated, given the clear divergence of interest?) Goldman beat the bushes, passed around the flipbook, as Jon W put it, they entered the realm of “sale/promotional”. Paulson could have expressed his view by calling dealers and buying protection on existing issues. As tom pointed out, he could have bought protection on the ABX. But since the market was in fact unbalanced — there were no unprodded longs to balance the short position he wanted. Market prices would have moved. In fact, they would have moved in the direction of price transparency, of accurately representing the lack of creditworthiness of RMBS assets. But they did not so move, because people are not perfectly rational and markets are not perfectly efficient and valuation is inherently uncertain, so “sale/promotional” — the flipbook — worked. The seller found his buyer without much moving the price, and we were all the worse for it.

    I think that if the SEC wanted to pursue it, there might be a case for market manipulation as well as fraud. It was plainly a purpose of this transaction to prevent Paulson’s trade from shifting prices (and to create securities with synthetic leverage — he could bet against a lot of notional super senior for relatively small fixed payments). Avoiding adverse price action is an understandable goal, and there are perfectly legitimate means of doing it when, ultimately, the trade flow is balanced at current prices. (For example, “dark pools” match large buyers with large sellers without moving spreads or revealing strategic information.) But techniques that prevent price moves when trade flow is in fact imbalanced at current prices by conjuring a party (share repurchaser, analyst-hyped dot com buyer, scare-monger rumor-spreader) under false pretenses to take the other side is bad.

    ACA, like Goldman, did not represent itself as somehow balancing the interests of longs and shorts. ACA Management might not have cared, but I suspect risk managers at ACA Capital would have had a problem with that had they known. As I said in the post, I wouldn’t be surprised to find that ACA Management engaged in some wink-wink funny business. But that doesn’t exonerate Goldman.

    Market making is central to Goldman’s case because their alleged role as a market maker is their primary rationale for not revealing Paulson’s identity. Underwritten and broken sales are not typically anonymous, although brokered sales can be if the parties are okay with working through blind agents. There’s a clear asymmetry with one party being anonymous and another party being known in a brokered sale, but if the disclosed party understood the arrangement and was okay with it, okay. But there were not two agents here, only one. That one agent did not plainly reveal the role of Paulson to e.g. IKB, did not represent this as a brokered sale, and did not indicate that the Goldman was acting as an agent for Paulson. By default, if this deal were a brokered deal (which would occasion two agents) or an underwriting (which it plainly was), Paulson would have had to be revealed. His name could have remained hidden if the longs were okay with that, but his existence and role could not have been. The only way to justify his disappearance from any formal communication is to say GS was a market maker and Paulson was just the party on whom the firm laid off risk, In that case, Goldman would have had no obligation to disclose. But it is economically untrue. Paulson was the initiator of the trade, the adverse selection risk was with the longs.

    BTW, we have to distinguish the IKB/ACA notes ($192M) and the later $909M wrap. Goldman was a straight market-maker with respect to the wrap: Paulson initiated the trade, Goldman laid much (but failed to lay all) of the risk off on ABN, who laid it off on ACA (Ha!). The only problem with that side of the transaction is that it was based on a structure that may have been biased in a manner that ACA Capital did not understand and that was in a certain sense the “poisoned fruit” of the original malarrangement. The $192M of notes were issued and the portfolio selected in a manner that violated Goldman’s duties as an underwriter.


  28. James Surowiecki writes:

    Steve, I think you’re misinterpreting what Ezra was saying. Ezra wasn’t talking about whether Goldman had a responsibility to disclose that Paulson was involved in the deal — I suspect he, like I, think that Goldman did have that responsibility, ethically and perhaps legally. What Ezra was talking about was the question that kept coming up in the Senate hearing, which was whether Goldman had a responsibility to disclose its own opinion of the deal (in other words, whether it thought this was a good deal for the buyers). And I don’t see where that responsibility comes from. You write, “They thought Goldman Sachs was working for them, underwriting securities of a special purpose entity it was putting together to satisfy investor interest.” But if an investment bank is underwriting an issue, it’s not “working for” the potential buyers. It’s working for the issuer, isn’t it? That doesn’t take away the issuer’s disclosure requirements (and Paulson’s involvement would, I think, be something that should have been disclosed). But do you also think that an underwriter is required to tell potential buyers (other, obviously, than clients to whom it owes fiduciary responsibilities) what it thinks of the deal, even if that would reduce the price its client (the issuer) would get?

  29. Steve Randy Waldman writes:

    Don & carping demon — I think you guys have this exactly right. The Supreme Court even agrees with you. From SEC v. Capital Gains Research Bureau, Inc., et al:

    In the light of the evident purpose of the [Investment Advisers Act of 1940] to substitute a philosophy of disclosure for the philosophy of caveat emptor, it cannot be assumed that the omission from the Act of a specific proscription against nondisclosure was intended to limit the application of the antifraud and anti-deceit provisions of the Act so as to render the Commission impotent to enjoin suppression of material facts.

    I’ve just begun to look at legal stuff; it was never really my interest. I don’t know whether this act would apply to Goldman in the current context. Was Goldman an advisor for the purposes of the act? It certainly is a “registered investment advisor”, but would this act apply to Goldman its role as underwriter or perfidious agent? I don’t know. The actual case was a front-running case: an advisor was purchasing shares for his own account before hyping them. That’s no so different from Goldman hyping assets that its favored client was hoping to sell. But again, I don’t know the applicability of this act, or whether it’s been superceded, etc.

    But the Supremes clearly state what Don and carping demon claim. Caveat emptor is not the law of the land. Note that the Supremes were not saying that investment advisors had a duty to recommend only products they knew would work out. That’s impossible. They are saying that in terms Jon W would agree with, of “selling exposures” accurately, executing fairly, disclosing completely.

    But to say that Goldman did not have to disclose, for example, because more diligent investors could have relentlessly inquired is unsupportable under the Supremes philosophy. (Again, I can’t comment now on the applicability of this language, more recent case law, etc. I mean this in a broad philosphical sense.)

  30. Steve Randy Waldman writes:

    Distant Observer — Excellent.

  31. […] this matter? To think about that, I recommend yet another post by Waldman, in which he takes apart Goldman’s claim that it was brokering a trade between a […]

  32. Steve Randy Waldman writes:

    Morgan_03 — In the “Deconstructing ABACUS” post, I pointed out that the way this deal worked was similar to a product called a “bespoke” or “single tranche” CDO. How “off the shelf” this product was I suppose depends in part on what Goldman and other players in the industry had on their shelves. I’m skeptical of a characterization as “off the shelf”, because the prospectus makes it very difficult to distinguish the product from a vanilla synthetic CDO — if it was familiar, I think they would have characterized it more clearly. But I could well be wrong about that.

    According to Sandrew a synthetic CDO should be termed bespoke if the reference portfolio is custom selected (rather than, say, a standard index) and single tranche if rather than the credit derivatives are written on tranches of the (notional) CDO, rather than on the underlying debt. Under that convention, ABACUS was both bespoke and single tranche. (Protection on three single tranches was sold here, though.)

  33. Jon W writes:


    On the topic of market makers, I know that they adjust based on trying to figure out order flow and the balance between buyers and sellers. The difference here is that market makers must put up a two sided quote bid /ask and size and are obligated to perform any trade that hits this quote. You and I in our trading are not. This is a pretty big difference in why order flow is very material to them. It is this obligationg to trade at the quote. (This is what I was referring to in my last comment that you linked, they cannot change their quote once they know you are the buyer, they can only change it after they have executed.) This is pretty important.

    On the issue of disclosure… As you say disclosure in offerings is necessary to inform the buyer of everything relevant to the business they are “investing” in. It is this business and everything necessary regarding it’s performance, not the performance of the security that is material. In a normal offering the issuers and their motivations are material because they ARE the business. The CEO, management team and executives of a company selling shares, ultimately have a fiduciary duty to the new shareholders. This is why their economic motivations are material.

    In ABACUS this cannot possibly be true. Paulson, nor ACA nor Goldman have any impact on the business. ABACUS could “exist” as a business without the involvement of Paulson. This is clearly not true in a normal corporation, the management team are part of the properties of the business. The business entity in ABACUS is simply a static group of reference securities whose performance neither Paulson, ACA nor Goldman have impact upon.

    I think the diffculty we’re having is that I believe material information (as it relates to offering disclosures) goes to define what the business is. You’re viewpoint is that material information is anything that may change the perceived value of the SECURITY.

    I agree that knowing paulson or any one else was intending to short a security would likely change perceived value of the security. This is obvious. It does not however change the qualities of the underlying business. There are innumerable pieces of information that would change someones perceived value of a stock or CDO note, how could you possibly ever account for them. What is the IQ of the CEO? Does the management team feel the stock market is overvalued? Does the management team intend to hold its stock for 2 years, 5 years, 10 years?

    How I see it is that Paulsons role is only in what the perceived value of the security is, and not what the business actually is. My argument isn’t that ACA or IKB wouldn’t want to know or find it useful. They very well may have. It may even be useful. To some this may matter, to others it obviously does not. As a buyside investor myself, I wouldn’t care. This is basically the issue. If the standard is anything that may change the price of a security then it would be impossible for any disclosure to be sufficient. Each person here could name 10 things they would want to know from counterpartiers, issuers, underwriters that are not currently disclosed in offerring prospectuses, but still ‘useful’ in figuring out a price.

    We will probably continue to disagree on this but this is the rationale behind my viewpoint.

  34. Steve Randy Waldman writes:

    James — You might be right that I’ve misinterpreted Ezra. It was the phrase “structuring a trade for two clients” that set me off. My beef is really with Goldman, who is acting as though it is a perfectly normal thing for two adversarial parties on equal footing to informally tweak a trade without any clear process for adjudicating disputes or even meeting to argue.

    The point about underwriting is much more subtle. Normally you’d be right, of course. The underwriter works for the issuer. Still, the underwriter’s relationship to potential buyers is very different from a market maker, who is basically an adversarial trader. An underwriter has a duty to investors, particularly to disclose all material facts and not to mislead. Ordinarily, the identity and economic position of the seller both before and after the transaction are material, disclosed, and subject to negotiation prior to the actual issue. Nearly always deals are structured to ensure that the seller retains a significant stake or has other exposure (e.g. via warranties and guarantees) to protect buyers from a lemons problem. There is no norm of anonymity in underwriting. On the contrary, the identity, intentions, and interests of the seller are crucial.

    But in this case, the issuer is a de novo entity. There is no preexisting seller: the investors are going to hold the first securities the SPV issues. Working for the issuer is the same as working for the investors.

    Now, since this is a synthetic CDO, there will be another class of claimant behind the noteholders in the prospectus. It is conceivable that one might consider a party that will take the short side of a derivatives portfolio to be in an economic sense the seller of the entity’s assets, and also a claimant or stakeholder in the entity, since that party will be due contingent cash flows. But that’s not at all obvious or normal.

    First let’s think about why. ABACUS was a little firm constructed to pursue an investment strategy. Suppose, as IKB and ACA presumably believed, there was no hidden counterparty and no predetermined structure. They were going to work with Goldman to put together an investment strategy was going to be tailored solely to the noteholders’ interests, the investors would deliver cash to the entity, which would then put together its intended portfolio. That portfolio might include cash bonds, various derivative positions, perhaps stock or lumber or whatever. With a varied portfolio determined by the interests of the founding investors (and not some investor who wished to dump), there would be no possibility of an “implicit seller”. The variegated positions would be purchased on the market, via Goldman as a market maker. In this case, it is pretty clear that there is no conflict of interest — the issuers and investors are the same, they are Goldman’s sole client, Goldman’s duty is to them.

    Now, in fact ABACUS was a synthetic CDO, one that to me seems pretty exotic. Does the picture change? Well, implicitly there are economic sellers of the credit, people on the short side of the CDS position. But in our free-form case, we planned to take derivative positions as well. That didn’t give counterparties any sort of control rights over our fund; that would pretty clearly be absurd. Now we will sell credit protection, buy some cash bonds, and enter into some miscellaneous derivatives positions intended to turn the cash bonds into synthetic risk free floaters. (That’s actually what ABACUS does.) Do we have any special reason to believe that the usually various parties on whom Goldman will lay off the credit risk should have any sort of control rights over the structure or command any loyalty from Goldman as underwriter? That is certainly not obvious. Why the credit sellers, and not the people with whom we entered into a swap to help synthesize the risk-free bond we needed? We would not, as a matter of general practice, expect that there would be only one seller of credit. Would Goldman owe some duty to the gaggle of investors on whom it would hedge its exposure from the deal? After all, it was very much not the norm for a market maker to find a single counterparty on whom the exposure would be hedged. And besides, we should be indifferent to how a market maker hedges. That’s why the market maker can keep the identity of such parties anonymous. So we are justified in assuming a black box behind our trades with Goldman that in no way impairs our interest in the new firm.

    So, again, if Goldman as underwriter is an agent of the issuer, for a de novo investment fund, there should be no conflict of interest. Goldman’s duty was to the noteholders, and solely to the noteholders.

    Now, it would certainly be possible to alter the terms of the arrangement with the knowledge and consent of all involved. Paulson could have established the fund as the founding investor, and had his hedge fund purchase credit protection from the SPV. Then he could have hired Goldman to underwrite a sale of notes. That would have been fine, from an ethical perspective.

    But it would not have worked as a practical matter. Paulson would not have been anonymous, his motives would be scrutinized. He would have been selling all or substantially all of his interest in the entity, which normally takes a deal out of the relatively cooperative realm of underwriting and into the hard-nosed, fully adversarial world of M&A. It would be obvious that Goldman was selling the credit portfolio that the enity synthesized, and purchasers would wonder why. Paulson would be asked and have to answer about his economic position, which would be revealed to be speculative (i.e. he wasn’t selling to hedge some new exposure). If he did not answer, the deal would not be done by an entity like IKB, which for all its yield-seeking CDO mania was not a hedge fund but a bank. IKB could only have proceeded with the repuational cover provided by Goldman and ACA. Paulson would have had to find a speculative hedge fund to take the other side of the trade, and (as Tourre’s e-mails remind us) they would extract a steep price.

    So, Goldman and Paulson hid there relationship, and ABACUS was formed as an entity whose owners were solely IKB and ACA, and Goldman’s duty as underwriter was soley to IKB and ACA.

  35. […] Steve Waldman pretty much sums up my views of this issue and why I think Goldman is ultimately culpable in the SEC civil suit against them: There is no circumstance where an investment bank “structures a trade for two clients” whose interests are opposed when the terms are anything but standard. I mean, really. Think about it. It’s Orwellian — Goldman calls a practice that is absurd on face “market making”, and suddenly it’s normal except for technical questions about who picked what securities or who should have suspected something. […]

  36. pebird writes:


    Your point on the janitor example made my conspiratorial mind start up. There are levels of complexity in this situation: 1) deal structure, composition, due diligence, disclosure – mechanics, 2) the relationships between seller/agent/buyer, 3) buyer/manager/owner.

    If I had good market information and thought I could structure something to protect myself while taking advantage of poor information/control owners have, then what I need to do is give the buyer/manager plausible deniability – I can leverage the buy side compensation/fiduciary conflicts and claim that only sophisticated investors were involved. Also, it is owners’ problem if their management is so easily bought off, and shows that owners s/b more actively involved. I have all the bases covered – simplistically speaking.

    I am surprised at the number of commenters who want to use the term “unethical”, as if that had some meaning here. In the context of upcoming financial reforms, we need to discuss what constitutes fraud – how much and about what can someone lie in order to close a transaction – it seems to be the consensus that a certain amount of salesmanship/lying always takes place (which is why ethics is a distraction) – but what are the boundaries of deception we are willing to tolerate?

  37. James Surowiecki writes:

    “So, again, if Goldman as underwriter is an agent of the issuer, for a de novo investment fund, there should be no conflict of interest. Goldman’s duty was to the noteholders, and solely to the noteholders.”

    Yeah, I don’t get this. I don’t understand how the absence of a real “issuer” converts Goldman’s duty to the issuer into a duty to the potential buyers.

    Regardless, I think there are actually three separate issues here that are being conflated. The first is whether Goldman had the responsibility to disclose Paulson’s role in helping pick the securities, because this was a material fact about the issuer. I think the answer is clearly yes (at least ethically). So in the specific case of Abacus, Goldman acted badly.

    The second is whether Goldman had the duty to disclose that it was going to be laying off the risk from the deal on Paulson. In other words, assume Paulson had no input into the actual CDO, ACA had actually picked the securities the way it was supposed to, and Paulson had then decided he was willing to short the securities. You seem to be arguing that even in this case, there was a duty to disclose that Paulson was on the other side of the trade, because having one person on the other end of a trade is different from having a gaggle of investors. I’m ambivalent on this question — it would be better were Goldman to disclose how it was planning to lay off its risk, but given that the fundamental value of the CDO is completely independent of Paulson’s actions (and in this was quite different from a stock or bond), it’s not obvious to me there’s an affirmative duty to disclose.

    The third issue is whether, even if there had been a gaggle of short investors (rather than just Paulson) on the other side of the deal, Goldman had a duty to tell potential buyers that it thought buying a subprime CDO in early 2007 was a bad idea. This is what Levin was arguing, and it’s this argument that Ezra says is a mistake — as do I. Basically, I think we should give up on the idea of investment banks underwriting deals in the sense of vouching for or giving their imprimatur to the deal — underwriting should be understood as simply structuring a deal on behalf of someone who wants to sell (equity, debt, whatever) and ensuring that all material facts are disclosed, but not as expressing any opinion as to whether it’s a good or bad deal.

  38. David Larsson writes:

    <blockquote="Goldman has the nerve to call both sides of the arrangement 'customers'."
    Thank you. This fundamental point seems to have become obscured by the squid ink that issued forth from Washington earlier this week.

    Consider what would happen to a legal advisor in an analogous situation — what if Goldman’s lawyers, rather than Goldman itself, undertook to represent all of (a) Goldman, (b) Seller, and (c) Buyers in this transaction? In most states, it could not do so unless Goldman’s lawyers “discussed this issue in advance, and secured the agreement of all clients for [Goldman’s lawyers] to remain in the matter” (New Jersey wouldn’t permit it at all, even after informed consent). “Discussing the issue in advance” would certainly include disclosing to the Buyers the fact that a “Seller” other than Goldman existed, and the identity of Paulson as the “Seller.”

    Now let’s suppose that Goldman’s legal advisors decide to move beyond the advisory role and become “market makers” by investing their own money in the deal on the Paulson/Seller/Goldman side, i.e., adverse to the Buyer’s interests: in most states, in addition to making full disclosure to the Buyers and obtaining written consent, the transaction and terms of the legal advisors’ investment would have to be “fair and reasonable” to the Buyers and “fully disclosed and transmitted in writing in a manner that can be reasonably understood by” the Buyers.

    Let’s suppose that Goldman’s legal advisors ignored all of these issues, and simply exercised their constitutional right to breach the rules of professional conduct by representing all parties without full disclosure and taking an undisclosed stake in the deal. The consequences? Goldman’s legal advisors would subject themselves to loss of their professional licenses and civil liability for malpractice on the basis of conflict of interest.

    Why ought society tolerate behavior from Goldman that it would not tolerate from Goldman’s legal advisors? After all, legal advisors do not have Goldman’s capacity to make — or break — markets.

  39. Michael writes:

    James/Steve–I agree 100% with James that the fault is not a failure to advise as to whether the investment was a good one but rather to provide full disclosure of material information, i.e., that the person who in effect selected the portfolio of assets did so with the specific intention of shorting against them. The purpose of the securities laws–which is what the SEC is enforcing–is one of disclosure quality, not investment quality.

  40. Steve Randy Waldman writes:

    James — The duty is still to the issuer. But the issuer and the buyers have no conflict of interest. There’s nothing much to get. Goldman is arranging a deal with its clients on one side and “the market” on the other, where the market is defined as what can be arranged via arms-length transactions. It gets unethical if there’s a hidden seller whose interests are somehow intermingled with the issuer.

    On your three issues:

    1) We agree on ABACUS and Goldman’s not-formally-disclosed active influence in selection. That’s the easy question.

    2) The reason Goldman has to disclose Paulson’s involvement is because Paulson is the initiator of the trade and that is material to its economics. Suppose Paulson goes to Goldman, and Paulson just describes the characteristics of the portfolio that he wants and waits. Goldman then markets a deal that fits Paulson’s requirements for the other side of the trade. Even though Paulson stays far away from the halls of ACA, he has implicitly selected both the composition and timing of the trade when he set the acceptable parameters. In geek speak, IKB now faces adverse selection risk. An investor, like an insurance company, faces adverse selection risk when a counterparty who might know something chooses the nature and timing of a deal. Like an insurance company, investors in that situation always jack up the price of their trade. Sure, it is possible that any given trade initiator doesn’t know anything, or an insurance customer is perfectly healthy. But in a statistical sense, when someone comes to you to sell, the stuff you’re buying is much more likely to be rotten than if you randomly timed the trade and selected from the complete universe of investments. Current market prices, if they’re at equilibrium, represent appropriate pricing for a random trader. If you take counterparty-initiated trades at market prices, the expected value of your trades will be negative. Over time and multiple trades you will be a sure loser.

    (This is why market makers set a spread. People who initiate buys pay a bit more, people who initiate sells recieve a bit less, than the midpoint “fair” price to compensate for adverse selection. It is also why the spread and price inevitably rise as the size of counterparty-initiated trades grow — that’s both evidence of more conviction and imposes greater risk on the responding trader.)

    By not telling our IKB stand-in that Paulson suggested the trade, it is as if Goldman is letting sick people come to it, and then organizing groups with a higher-than-normal frequency of sick people and marketing them as randomly selected pools. It is a subtle way of screwing the insurance company.

    When an investment bank markets a product, what should its duty be to the clients that accept? Fiduciary may be too high a standard. But knowingly marketing negative expected value trades is too low. In any given trade, the IB may sincerely believe that the initiator (e.g. Paulson) is dumb, ill-informed, a rube. But unless you posit perfectly efficient markets or you make heroic assumptions about how dumb and ill-informed the people who would initiate bespoke deals are, then on average, clients responding to counterparty-initiated marketing campaigns are going to lose unless they insist on adverse selection premium.

    [If you are very attentive, you may object that I seem to be both assuming and dismissing market efficiency here. I’m not — I am assuming market prices are unbiased but variable estimators, so that on average “prices are right”, but at a given time, any price can be high or low, which mispricing a savvy or informed trader can exploit.]

    Note: Goldman would have no duty to disclose anything about Paulson if his involvement only appeared after the deal had closed. Then Goldman is just making a market then, and happens to find a counterparty on whom to easily lay off the risk. The problem is Paulson’s implicit role in selecting both the composition and timing of Goldman’s marketing campaign if he is the impetus and initiator of the deal (how ever passive he may be while it is implemented).

    3) If Goldman markets a deal, however it intends to lay off the risk, it shouldn’t believe the deal to have negative expected value before fees. That is my standard — when Goldman is the initiator, its duty to clients is that the deals it markets should perform on average over time at least as well as rolling T bills. Obviously that’s hard to measure and enforce, but in principle I think it’s right.

    However, if a client comes to Goldman and asks the firm to make a market for it, it doesn’t matter how Goldman intends to lay off the risk (or even if Goldman actually wants the other side of the trade). Goldman has no obligation whatsoever to disclose what it thinks of the client request. Market-making is a trading strategy. It’s arms-length and inherently adversarial. There are rules that pertain to market makers in organized markets, but those are choices made by exchanges. If a customer comes to Goldman and asks for a complex derivative trade, Goldman will offer a price, which will undoubtedly (and properly) include an adverse selection premium, and also a nice profit margin. If the customer doesn’t like that price, she can call up JP Morgan and look for a better price. Client-initiated arms-length trades are a game played by grown-ups in a competitive marketplace. We might wish to make that marketplace more competitive (e.g. force derivatives to exchanges, or enforce antitrust if we think the dealers are colluding). But the dealer’s obligation is simply to offer a price and take the trade if the client wants it, full stop. “Client” is really the wrong word to use for unmarketed, customer-initiated trades. The customer is simply a counterparty.

    Goldman owes a duty to a client if the client hires the firm as an advisor, or if Goldman itself markets a deal to investors. Goldman can initiate trades with market-makers and arms length counterparties as it pleases. Goldman can also develop menus of always-available trading products and send them around without either touting or discouraging them. Customers can then take the exposures they want, when they want, without imposing any duty onto Goldman. But if Goldman markets a deal, encourages investors to invest at a certain time and in a certain product, then its duty is to offer what it believes to be at least an actuarially fair deals.

    You hit on this in your (excellent) live chat yesterday. There’s nothing wrong with being a bookmaker. In fact, we depend on bookmakers so that we can make the deals we want when we want, and to get market information into prices. But you are no longer just a bookmaker if you tout specific bets at specific times. That opens up lots of scope for corruption, and it screws up market pricing. If investment banks want to play that game, they’ve got to hew to standards that limit the potential for abuse.

  41. Steve Randy Waldman writes:

    David — Great point. I’m not sure about who’d win an unpopularity contest, but the lawyers seem to have one up on bankers for ethics, at least in theory…

  42. Steve Randy Waldman writes:

    Michael — I don’t disagree, but I think that if investment banks market products in response to signals that imply (in a statistical sense) underperformance for the product (e.g. Paulson defines the general terms and timing of a deal and then stands aside while Goldman touts it), that is a disclosable material fact. A reasonable investor would want to know, and a reasonable investor would want to include an adverse selection premium in her pricing.

  43. JKH writes:


    Far left field question:

    how does the buyer of the super senior tranche assess the credit risk on it?

    there’s the risk of the underlying mortgage performance, but presumably that’s compounded by contingent credit risk – how should that have worked in this case?

    Don’t know if its relevant; just interested.

  44. Steve Randy Waldman writes:

    JKH — You couldn’t be irrelevant or uninteresting if you tried.

    I don’t know — they modeled that stuff, but I don’t believe the models. If you think you can price the tranches below, you can bound the value of the super-senior stuff, it’s got be more than the not-so-super senior. (I think… but then swap spreads were always suppose to be positive.

    I’d defer a bit to James’ now out of fashion “Wisdom of Crowds”. Markets may be inefficient, but i think markets dominated by professionals (read agents with distorting incentives and “standards of practice” that force herding) are worse than markets open to all. If you asked me how to price ’em, I’d give every tranche some kind of a ticker symbol, cut them into $10 par value pieces, put everything known about ’em on the web, and make sure every punter has access to ’em on eTrade etc. I’m sure they’d be mispriced, but I suspect they’d have been far less mispriced in early 2007 than they were when valuation was left to professionals.

  45. Michael writes:

    Steve–Your geek speak is a bit beyond my reading level. But if you’re saying that it should have been disclosed that the securities selected were selected specifically because they were believed to have a high likelihood of default, then we’re in full agreement. Goldman’s trying to obfuscate that issue by saying that all subprime loans are by definition expected to perform poorly, and that that was disclosed to and known by investors. But, to use Tom Montag/Senator Levin’s word, if you’re trying to sell something that is specifically designed to be the shittiest of the shit, then you need to disclose that, too.

  46. James Surowiecki writes:

    Is the Wisdom of Crowds out of fashion? Now you tell me.

    “But you are no longer just a bookmaker if you tout specific bets at specific times.”

    I think this is really the crux of the issue, because it’s not clear to me that it’s true — any casino sportsbook comes up with a whole host of prop bets for, say, the Super Bowl that it wouldn’t normally offer and that are typically negative expected value for bettors (relative not just to doing nothing, but relative to betting on the point spread). These bets are initiated by the sportsbooks, but it seems like it would be odd (that is, wrong) to demand that the sportsbooks tell their customers “your chances of making money on this bet are not very good.”

    Now, we can say that this isn’t the same as having a Goldman salesman call up an investor to try to sweet-talk him into buying a deal — and I agree it’s not exactly the same. But ultimately I think we would be better off if people understood the two things as the same, and trusted the Goldman salesman as much (which is to say, as little), as they trust the sportsbook. I think things would work better if, at least with regard to the trading of financial instruments, we replaced the idea of “clients” with “customers,” and abandoned the notion that investment banks can be relied on to vouch for the value, in any sense, of an asset. Again, this doesn’t eliminate the need for disclosure of material facts, but effectively I just think all investors should assume that investment banks are not vouching for anything about the deal other than that the relevant material facts have been disclosed, making investment banks’ opinions as to the expected value of deals irrelevant. And if that ends up making investment banks themselves irrelevant — or changes their rule into being more explicitly bookies — I think that would be a net gain for the markets.

  47. MR writes:

    JKH/Steve – There were only two buyers/insurers of the super-senior tranches – the banks themselves and the monoline insurers/AIG. The insurers, believe it or not, looked at the seniority of the tranche and concluded it was “zero-risk”. Some of the super-seniors at the peak of the boom were insured for less than 10 bps running.

    If you ask any CDO biz head in a bank, they’ll blame the models as well for their outsized super-senior exposure. But I don’t buy that for a second – the poor models were just a tool to hide the true nature of the risk. Remember that even in synthetic CDOs constructed with a portfolio of CDS instead of a bond, someone had to take on the super-senior exposure else there was no deal. So CDO desks tried to justify holding onto the super-senior even in an unhedged manner to keep the deal-flow going. And when they did hedge it, they used cosmetic methods that just swept the risk under the carpet. UBS had a strategy called AMPS which was like the extreme opposite of the Magnetar trade – essentially long super-senior and short the underlying portfolio in 2-4% of the super-senior notional, and voila the risk system shows the portfolio as zero VaR! riskless profits! Even allowing for faulty models, I find it hard to believe that the CDO desk didn’t appreciate the existence of higher-order risks in a delta-hedged portfolio.

    What enabled the banks to hold onto the super-senior when no one else could is the cheap leverage they could get – on a 40-1 leverage, even a 20 bps running spread can add up to a very high return on equity.

    UBS came up with a detailed look at why they lost all that money in a shareholder report here http://www.ubs.com/1/ShowMedia/investors/releases?contentId=140331&name=080418ShareholderReport.pdf . Its a bit heavy on the jargon but I’ve analysed some of the relevant parts at the end of this post http://www.macroresilience.com/2009/11/06/a-rational-explanation-of-the-financial-crisis/

  48. JP writes:

    With respect to the real estate analogy — there are actually situations where one person serves as a “dual agent” on both sides of a transaction (legal in some states). However, in some states, one side is a client (the seller); the other side of the deal is a customer (the buyer). Under these circumstances the agent must disclose to the customer that his or her fiduciary responsibility rests primarily on one side of the transaction. In other states the dual agent must treat both sides “equally” and cannot provide advice to either side with respect to terms, conditions, and price. This is one reason that most agents tend to prefer to execute a transaction as a sole agent on just one side of the transaction (so as to limit liability and potential conflicts of interests). I’m not an expert on securities laws, so I don’t know if this analogy is even applicable in this particular case. However, as a general rule it seems disclosures about relationships in a deal are just as material as disclosures about the underlying content of a deal. In the Goldman case there is an issue of an alleged misrepresentation (e.g. suggesting that Tourre told ACA that Paulson was a long investor — when in fact he was not); there also appear to be issues centering around alleged omissions of material fact.

  49. Alex writes:

    How come ACA hasn’t been heard from much? Shouldn’t they be more pissed off or suing?

  50. JKH writes:


    I was thinking of something a little different in my rather vague question. It goes something like this:

    Consider for purposes of illustration the very extreme polar counterfactual, whereby the underlying mortgages actually perform, but either or both of Goldman and Paulson go under.

    Notwithstanding the miniscule probability of such an outcome, the CDO super senior “investor” (i.e. ACA) should incorporate it (in theory) in his credit analysis at the outset. The consequence of such an outcome as above, I believe, is that even though the underlying cash mortgages perform, the supersenior investor loses the income from the synthetic cash flow, because the CDO no longer has the income from the counterparty on the short side (I think).

    i.e. the full (in theory) credit risk analysis for the supersenior investor involves mortgage credit analysis plus counterparty credit risk (as in any swap).

    So I was trying to figure out how the CDO supersenior investor should analyze counterparty credit risk, at least in theory (and I know these things were considered better than AAA), and whether that at least theoretical analysis has any implications for information that would be required (again in theory) about Paulson as a short counterparty to the CDO. I’m not clear on where the counterparty risk to the SS investor would be in theory – Goldman or Paulson or both – and its clouded by the fact that supersenior was considered better than AAA – which makes it highly theoretical, I guess.

    This is a real tangent, and I don’t know if it makes any sense. But somewhere I see a potential connection to information requirements re Paulson – or not (arguing here against my own case in general).

  51. rootless_e writes:

    “How come ACA hasn’t been heard from much? Shouldn’t they be more pissed off or suing?”

    “Hey we are suing you for correctly assuming we were too lazy to do due diligence and too confident of our ridiculous model to consider if it was still valid even after all the smart money was not only betting against us but giving talks about what saps we were” – maybe not a good case.

  52. Again I agree with MR. The models are smokescreens to conceal the fact no one knows that the tranches are worth.

  53. Kirill Leonevich writes:

    Mr. Waldman, thank you for your superbly detailed and nuanced analysis. I keenly appreciate the distinctions that you draw between the duties of broker, market maker and underwriter. That said, I would argue that the distinction is unnecessary to find misfeasance in the ABACUS affair.

    Let us use the bookie standard as our starting point. (As a tribute to Sen. McCaskill, if you will.) This, I believe, is the standard that Mr. Blankfein proposes – to wit, that Goldmans job is to make sure that the products that they sell do what they are supposed to do. Or, if I may, that if they offer a punt on Dick’s Last Stand winning the 4th at Belmont, that it will indeed be a punt on Dick’s Last Stand winning the 4th at Belmont. What horse they think will win the race, or what horse in fact does in the race, ought not to be their concern.

    I wholly agree with this.

    Now, this begs the question, what are the material attributes of the product that Goldmans sold – the attributes that the product must in fact possess, if Goldmans standard of conduct be upheld. One such attribute, as for example per the pitchbook, is that the reference portfolio was selected by an expert advisor using some methodology. Now, what that should mean to the prospective investor is, I argue, suspiciously vague. It’s a bit like an advert showing a chap drinking a Brand X lager whilst nubile lasses (or, depending on where published, nubile lads) watch with doe-eyed lust. Does that imply that my drinking a Brand X lager will bring me such advantage? But let us put aside that matter, and suppose that a prospective investor would assume potential alpha – superior performance (with respect to a randomly chosen pool of nominally comparable assets) consequent of discerning selection. Discerning selection – performed with intent to benefit the investor – is the essential feature. Goldmans represented that as an attribute of their product. They were advertising alpha.

    Now, surely, a short may be a discerning chooser. (Indeed, one reason why I consider shorts “socially useful” is that I hold them on the whole to do better research than longs, and therefore to contribute much to price discovery. I admit that this is a gross generalisation, which I propose without evidence.) Nevertheless, I take it as a principle of agency that one ought to disclose conflicts of interest – incentives that may undermine the zeal with which one discharges one’s commission. Whatever Paulson’s expertise, his involvement in establishing the reference portfolio is surely not the discerning selection that a prospective investor would reasonably have expected, given Goldmans’ description of the deal. This is not a matter of hype in a sales pitch. It is a matter of materially misrepresenting the attributes of a deal – a matter of fraud.

    Selling alpha is dodgy business. Those who tout their scrupulous agency as do Goldmans should sell alpha with utmost care. There is little moral ground between alternatives brokers who tout their prowess in selecting managers, only to send all their clients’ money to Bernie Madoff, and brokers of bespoke deals offering exposure to expertly chosen assets, which were expertly chosen by a short. Whether Goldmans were acting as market maker, broker, or underwriter for the ABACUS deal, they misrepresented what they were selling.

  54. A Reader writes:

    See page 25 of the prospectus and then see Fab’s testimony over who knew what.

  55. […] reading some of the crazy shit that state elected GOP officials are doing, Steve Waldman’s excellent explanation of the Godman Sachs trading that caused controversy, and Ted Cox’s undercover experience at a […]

  56. MR writes:

    JKH – I get your point. Usually such counterparty credit issues are taken care of by having a collateral agreement where one counterparty posts collateral to the extent that the mark-to-market is against him. But the whole point of ACA’s business, much like Berkshire Hathaway’s strategy with selling long-dated equity put options, was not to post collateral and only have to pay ultimate losses. This exposed both counterparties to credit risk – I’m not sure ACA ever gave much thought to it given as you said the remote nature of their risk. Goldman obviously cared which is why they entered into another overlay hedge with ABN/RBS.

  57. JLF writes:

    JKH – GS intermediates the trade so only GS has an exposure to Paulson which presumably is managed through a margin arrangement. Note investors at supersenior and other levels have minimal exposure to GS as a GS default would be a mandatory redemption event, i.e. the collateral would be liquidated to repay the note principal (see table on page 6 of OC). These structures were specifically designed to remove counterparty risk, hence the collateral.

  58. MR writes:

    And IKB/ACA only face Goldman here so Paulson’s involvement as Goldman’s hedge is not relevant to them from that standpoint.Even if Paulson goes bust, Goldman still owes the SPV/IKB and ACA.

  59. JLF writes:

    If there had been note exposure to GS or ACA or Paulson counterparty risk the ratings would have been limited well below AAA/Aaa.

  60. JLF writes:

    Well I guess to be technically correct IKB/ACA as note investors only face the SPV which then faces GS which then faces Paulson. Key point is that collateral protects note investors from third party counterparty risk. In case of Paulson bankruptcy GS still owes the SPV and so GS will have been very keen to manage that exposure through a margin arrangement. Exposure wouldn’t have been huge relative to note balance as only risk is that Paulson can’t pay CDS premia.

  61. farang writes:

    Okay, enough gibberish. Goldman Sachs and Paulson put together a shit sandwich, and sold it to another client as Filet Mignon.

    As I wrote in comments on another blog (and oddly enough, Tabibbi has almost the exact same anaolgy two days later…)it comes down to this:

    Do we really want a world, a nation, where every little thing we do requires “due diligence? Really?

    So tomorrow I awake, and go to brush my teeth: out of paste. I go to store to buy some: do I need to ask for a sample to have tested to see if it will poison me? Do I need to run a check to see if the manufacturer and store owner have insured me from death against poisoning, naming themselves as beneficiaries? So, I buy the toothpaste, and head home…and my brakes start to squeek. Do I then call Better Business Bureau, to see if there are any reputable brake repair shops? Do I also do a background check on the BBB officers, to be sure they don’t own any brake shops?

    I decide to go to Goldmen-Stops brake shop. They knowingly sell me defective brake pads they have bought at deep discount, but represent it to me as 5 star quality pads. They put a sign saying I cannot enter shop area, and hire a drunk from the alley behind shop to install my brakes, then call their buddies a AUG insurance too insure me and my family against accidental death from brake failure.

    THAT is the cogent point: who the hell would EVER do “business’ with these scumbag conmen again, except OTHER CON ARTISTS.

    HOW MANY of these type deals are done everyday? And, best part? When it all goes to shit, they bribe the authorities through lobbyists to bail them out.

    ANYONE defending these type scams and cons are criminals too. ANYONE.

  62. JKH writes:


    Thanks. One point I wasn’t recalling clearly is that is that Goldman obviously stands between the investor and Paulson. In a way, that’s a marginal point in favour of the view that the investor didn’t need information on Paulson, although probably a minor one (sorry Steve).

    I’m aware of the collateral mechanism, but the issue still isn’t completely clear to me.

    I believe the purpose of collateral isn’t to protect the investor in this case – it’s to protect a money position for the shorts on the CDO, under an obligation to the longs (who are short the option). Posting collateral is a cost to the investor here – i.e. a liquidity cost – so recovering it can’t be a net benefit -especially from the perspective I was referring to.

    I was interested in the contingent credit risk on the deal for the investors – which is their counterparty credit risk when the primary credit (mortgages) ends up being good but the counterparty risk (which you straightened out for me as Goldman) goes bad.

    That contingent credit risk exposure is the income that the CDO owes the investor as earnings on the long position – i.e. the CDO fails to deliver its income obligation to the investor according to the synthetic contract requirements, when the mortgages are performing.

    The best the investor can do is call back all his collateral if that contractual liability has been neutralized in a good mortgage market – in the scenario I describe, the shorts are no longer in the money, and really have no economic claim/right to the collateral – unless its written into the contract as overcollateralization, and in that event, the investors’ recovery of previously posted collateral might even be threatened in the scenario I describe, but that’s really stretching my point.

    MR, if I’m reading you correctly, you’ve alluded to all this with the suggestion that the investor might actually demand net collateral in such a case. That hadn’t occured to me, because mortgages don’t actually have to be “in the money” for the investor to face the scenario I describe.

    “Blah” as SRW would say.

    Anyway, my point as intended is pretty far fetched in terms of practical significance – it requires a collapse of Goldman while the mortgage market flourishes – and we all know that could never happen. The VAR would never permit it.


    VAR – “a theory of ocean waves whose swells are prohibited from exceeding six feet”


    Thanks again for your responses.

  63. Danny Black writes:

    Maybe this is just confusing for people who don’t work in the markets. I can give you a direct analogy based on your housing.

    I had a flat in the UK which i sold in 2006 to a buy to let “investor”. One of the reasons i sold was because I was permanantly moving abroad but also I considered the house prices to be insane and bound to come down. I filled out a disclosure form about any flaws in the property – there were none that i knew of – and any material information – as a top floor flat the owner was responsible for paying for the new roof. However, I didn’t tell the guy that I thought he was an idiot for buying at the price he bought at and that in my view the prices had only one way to go which was down – I would be wrong for another 12 months, they shot up. As an “investor”, my opinion should not be material to him because he has a view on what the property is worth. As a trader, maybe you could argue that he would act differently if he knew i thought the price was over the odds because then he could screw me down on price. Now we now know that prices DID go down and belatedly I was right and he was wrong. But at the time of the sale, he was Mr brilliant property investor – I know because he told me – and I was an idiot for selling – which he also told me.

    I simply can’t see how Paulson’s view on the portfolio is material to someone who is investing in these products. Paulson had a clearly and publicly articulated view on the market. The reason GS keeps bringing up that that ACA and IKB were “sophiscated” is because as “experts” in this field they should know how to do proper due diligence by themselves and not be just blowing in the wind with other people’s views.

    I suspect that the SEC KNOW Paulson told ACA they were short. I suspect that the case will be shown to clearly have zero merit. I also suspect that is not the point. The point was to have a public whipping boy whilst the Democrats ram a half-arsed regulation bill through. The point is also to embarass GS into settling based on some tiny technicality that no one will understand but will be portrayed as “GS screwing their clients” not least by GS competitors.

  64. Danny Black writes:

    farang, apparently what you want is a world where you are not responsible for any of your decisions. If you decide that toothpaste is too expensive and brushing your teeth with rat poison is better use of your money then in your world there should be someone to physically stop you being a fuckwit.

  65. Danny Black writes:

    Kirill Leonevich, except the portfolio wasn’t chosen by the short, expertly or otherwise. They were chosen by ACA, expertly or not.

  66. Jon W writes:

    I’ll try one last way to describe my thoughts regarding materiality. My basic argument isn’t that knowing paulson’s role or if goldman is short isn’t useful. Of course it is. The problem is that making it part of a standard of disclosure makes the standard impossible to fulfill and therefore it can’t and shoudln’t be.

    A much more philosophical construction. Function versus purpose.

    Function is the properties of an item as it relates to what it does and is.

    What are the functions of a ferrari?

    If you press down on a gas pedal the wheels spin. It is constructed from metal. Turn the steering wheel left and the tires turn left. It is red. It can go from 0-60 in 3 seconds… Each property above is quantifiable/verifiable by any party. These functions are the same and as true for you, for me, or even a cow you might run into on the road. Function is absolute and the same in all frames of reference. It describes who, what, where, when and how.

    The purpose of something on the other hand is the narrative applied to items in the world. It is an entirely human construct.

    What is the purpose of a shoe?

    For most people a shoe protects your feet while you walk. For others a shoe is a fashion accessory or a collectors item. For a dog it’s a chew toy. For the bacteria living in the sole, the shoe is home. Purpose is relative. It is frame of reference dependent. It is different for everyone. The purpose of an item describes Why it is. That we’re having this discussion illustrates that why as it applies to the item is impossible to define. Purpose does not exist without the person. Statements of purpose dont describe the item, they describe the person using the item.

    Look at the difference in sentence construction between the ones under function vs the ones under purpose. In “function” the subject in the sentences is the item. The sentences tell the story of the car. In “purpose” the subject in the sentence is not the item. The item is the object here. As statements of fact these sentences describe you, me, the dog, and the bacteria.

    The argument that KD, I and others are making is that since function is absolute, it can and must be the only way we evaluate materiality. The subject of disclosure is the item.

    Setting the standards of disclosure as an item’s purpose is to make it relative, making the subject all the people using it and therefore impossible to accomplish. This is true even if the great majority of the population shares the same why. Purpose is never a shared statement of fact.

    The functions of a security, what a security is, is a representation of assets. So long as the disclosures are a truthful representation of the assets, then materiality and the responsibility of an underwriter is fulfilled. While Paulson thinks the assets will go down is a fact, it is a fact about paulson and not the assets nor its representative security.

    I understand why this discussion is so difficult. Humans have a propensity for applying narratives and purpose to everything. We then believe that these purposes are elements of the world, rather than elements of us. We are interested in the human aspect and perspective of the world. While we may find it of little import, these perspectives still never change what the world is.

    Thousands of years ago, humans built a large ring of giant stones was built in Northern Ireland. This is stonehenge. In the absense of its builders everyone can still understand what was built. To this day no one really understands why it was built. We debate, and research and study ancient relics not because we want to know about the relics. Stone henge is and forever will be a pile of rocks. We pursue these studies to understand the people. It is the interest in other humans.
    Knowing why stonehenge was built doesnt tell us about stonehenge, it tells us something about the people who built it.

    If investors in a CDO like abacus were net short, and bought notes for the purposes of netting out their short exposure, paulsons role now is of no importance. These investors are now price agnostic, the purpose of the transaction is now to simply to net out risk. In portfolio management this happens all the time. The firms I’ve worked at have bought ETFs, CDOs options not to make money but to simply net out risk as it correlates to short positions in our portfolio. Now instead of worrying about whether the securities will go up or down, GS must disclose correlations vs portfolios, sectors and industries. After all if as the buyer I am wrong about how it correlates, then it does not serve my purpose and we’re in teh same bind again. See how this is an impossible standard to apply?

  67. Danny Black writes:

    Jon W, I think also that there is a hefty post-hoc analysis working here. Because Paulson HOPED it would decline in value and then it did then magically becomes something that somehow was ALWAYS going to go to zero.

    Personally, I believe this is a hefty reason we keep having these bubbles because when they crash the – in retrospect – idiots who bought – in retrospect – the crap get to play the victims. Rather than try and tie up the supply side in regulatory knots surely it would make sense to deal with the demand and one way to do that is to force people to take their losses. I remember reading in Singapore when people were demanding they be made whole on “risk free” lehmans structured notes that the PM said that anyone who was getting an above risk-free return MUST have understood he was taking a risk and therefore had no right to complain.

  68. carping demon writes:

    “I understand why this discussion is so difficult. Humans have a propensity for applying narratives and purpose to everything. We then believe that these purposes are elements of the world, rather than elements of us. We are interested in the human aspect and perspective of the world. While we may find it of little import, these perspectives still never change what the world is.”

    This would apply to a point-in-time snapshot of the world. But since humans exist only as part of the world, beliefs alter the world in real time. The frame is constantly moving, and real world is different each instant because of that. Humans don’t act against a backdrop of the real world, they are an integral and inseparable part of the real world…so these perspectives constantly change what the real world is. This is not to say that humans “make the world up”, just that the real world is inextricably bound up with human beliefs, just as it is bound up with the atmosphere. The inappropriateness of human beliefs is a matter of not being able to forsee the future.

  69. […] Randy Waldman breaks down the Goldman Sachs fraudulent deal- […]

  70. Philip writes:

    I’m new to this site (via Ezra Klein’s tout today) and have very thin credentials for participating here, but having just finished 13 Bankers and The Big Short, I’m motivated to learn and this looks like a great place to do it.

    I’m now persuaded that Goldman’s actions on the Paulson deal were unethical if not illegal. What I don’t understand is why Goldman would take this reputational risk in return for a measly fee. OK, since the Paulson deal was novel, maybe they saw an opportunity to do a series of deals like this one and therefore there was more than one fee at stake. But still, what a risk to run. (In The Greatest Trade Ever, Zuckerman says two other firms–Bear Stearn was one–turned Paulson down due to the ethics/reputation issue.)

    So that leads to a second question: if Goldman would risk their reputation for a fee (or stream of fees), what would they do if the firm itself were at stake? For example, when late in the game Goldman woke up to the exposure represented by their long-side subprime investments.

    Over a very short period of time Goldman appears to have moved from a strong net long position in its own subprime holdings to a net short position sufficient to reap a windfall from the subprime whirlwind it had sown. In order to pull this off without inciting a rush for the exits, I presume Goldman would have had to do this quietly.

    Is it reasonable to suspect that Goldman, as it unwound its long position, might have continued its marketing/promotional underwriting activities on the long side in an extended pump and dump campaign? Syncing up the timing of Goldman’s positional shift with its long side marketing activities would shed light on this question.

    Come to think of it, is it reasonable to suspect that Goldman dumped some of its long holdings on its own clients?

    Finally, is it reasonable to suspect that Goldman, in order to avoid moving prices in the relatively small short market, would have continued pushing long side investments in order to create new short opportunities without having to purchase existing short contracts?

    Any of these practices would seem to make the ethical/legal issues inherent in Goldman’s actions in the Paulson deal pale in comparison.

    Or am I taking all this too far?

  71. […] more here: interfluidity » A knife fight is not a mediation If you enjoyed this article please consider sharing […]

  72. This excellent thread has outrun me, but I do want to emphasize that you can tell a lot about people and their behavior by the way and manner of expression that they defend themselves. It is certainly possible that a law has been passed that allows for iffy behavior in my view. If that’s the case, looking at the basics can allow me to identify and judge that iffy behavior. Honestly, I often miss it. In judging a business, it is perfectly acceptable to look at its practices and see how it comports with my own values. That’s the kind of full disclosure I want. It’s the iffy areas that are especially revealing of motives and values.

  73. Philip writes:

    Danny Black:

    In the case of Paulson’s involvement in the ABACUS deal, why doesn’t your housing analogy translate as follows:

    Paulson approaches a housing developer (GS) and proposes to pay him a fee to build a house to be marketed and sold to a third-party, on the condition that Paulson is allowed to choose some of the building materials used by the developers’s builder (ACA). Paulson also informs the developer that he intends to buy insurance (CDS) on the house once it’s finished.

    Paulson proceeds to presents a carefully selected list of substandard and even defective materials to the subcontractor who rejects some but accepts much of it, and builds the house. Because Paulson knows vulnerabilities in the home inspection process, he is able to select materials unlikely to be identified by the inspectors (Moody’s), and the house passes inspection.

    The developer sells the house without disclosing his arrangement with Paulson or Paulson’s plan to buy insurance on the house. The buyer (IKB) moves in and the house burns to the gound six months later due to a fire caused by substandard wiring.

    Does the developer bear any ethical or legal responsibility for the loss suffered by the buyer?

    Does fault lie with the buyer because he (1) didn’t happen to think to ask about any hidden parties to the deal, (2) placed some degree of confidence in the integrity of the home inspectors and didn’t conduct his own slab-to-weathervane assessment of the property, and (3) counted on the value of the developer’s sterling reputation to restrain the developer from seeling a shoddy product.

  74. Danny Black writes:

    Philip, that isn’t what happened. The builder got to choose the materials, Paulson suggested alot of materials he HOPED would turn out to be defective. The builder accepted some of the materials and rejected others. The builder accepted some of them because unlike Paulson he believed the materials were good based on his many, many years of experience in building unlike Paulson who was a newcomer to the field. We NOW know that the builder was wrong and Paulson was right. We didn’t know that at the time and in fact the builder had got very rich – both personally and his company – by using the exact same sort of materials for the exact same sort of housing. It happened that in this case the house fell down whereas in the past they hadn’t.

    The question here is why the developer’s OPINION of Paulson’s OPINION is material as to whether the house would be sound or not. Why do we value Paulson’s view over ACA’s? Because we NOW know that Paulson turned out to be right and ACA wrong. Now unless you are suggesting that before selling any security or financial product that GS needs to build a time machine to find out whose view is correct I can’t see why you think Paulson’s well-documented view in 2007 should be a key factor.

    This was the point i was trying to make earlier about WHY this particular case was picked. Unpopular bank and a guy famous for making a right BET on the future. Easy to try and present that as a case of the unpopular bank selling a product designed to fail – which by the way ISN’T the SEC’s allegation – and so to give undue weight to the opinion of a guy that two years later would prove to be spectacularly and – more importantly – very publicly right.

    ( I am of course ignoring the news that apparently the SEC are aware of the fact Paulson’s trader TOLD ACA that he was going short ).

  75. Philip writes:


    Thanks for your reply.

    Well, where I think we’ll have to agree to disagree is that Paulson KNEW he had chosen defective materials though of course he didn’t know for certain the house would go up in flames. Still, he knew the deal was a crap shoot, though a very attractive one, especially after he had loaded the dice. But as we’ve known at least since Heisenberg, there’s no such thing as a sure thing.

    I think you’re mistaken in saying that the ethics/legality issue centers on Goldman’s failure to disclose Paulson’s OPINION. I agree that Paulson’s opinion was irrelevant. What IS relevant, and material, is Goldman’s failure to disclose that (1) it was underwriting ABACUS at the behest of another party which was hidden from IKB, (2) the hidden party had a significant role in selecting the mortgages contained in ABACUS, and (3) the hidden party was betting against ABACUS and therefore had presumably selected mortgages that would serve that end.

    You will respond that Goldman met its disclosure obligations with ACA, but as Steve argues persuasively, telling the “janitor” at ACA Management is not a suitable substitute for a formal, written disclosure to ACA Capital’s risk management officer.

    In any case, I suspect that you, like me, would be looking for someone to suffer if they had loaded your new house with faulty wiring, took a big insurance policy on the house and then collected $1 billion when it burned down 6 months later.

  76. Danny Black writes:

    Philip, the SEC’s case is that Paulson’s OPINION is material to the investment and that by failing to disclose that opinion GS committed fraud. If Paulson’s OPINION is material – and i think virtually ever market participent in any form who is not talking his book agrees it isn’t – AND if GS failed to disclose this – and it is then MAYBE it is fraud. I am not a lawyer but it would seem some form of intentionality would be necessary and possibly they need to show that GS hid it rather than neglected to disclose but that may not be the case, hence why the SEC could win on a technicality.

    The fact is that this case is being represented exactly how you think it is. That is that GS intentionally sold a lemon to an investor. That is NOT the case being brought. Or that it somehow hoodwinked an investor into buying something they didn’t understand. That is also NOT the case being brought. Or that GS somehow KNEW these products were not worth what they were or that ACA or IKB were some innocent babes in the wood buying something they couldn’t understand or that GS made massive profits off a collapsing subprime market. Also not the case being brought. Which is why GS has gone to the effort of not refuting the case – which I think clearly has zero merit – but refuting the presentation of the case.

    This is why the whole thing smells of politics. I suspect when the whole case comes to light the following turn out to be true:

    1) That ACA and possibly IKB knew about Paulson’s involvement and motives – this i think is a slam dunk prediction.
    2) That ACA and IKB thought THEY were the ones that were going take Paulson to the cleaners
    3) The SEC knew beforehand that the case had zero merit and made no attempt to settle
    4) That the SEC’s strategy consists soley of releasing lots of embarassing emails inducing GS to settle just to kill the bad news and hoping that when it comes to trial that the 12 peers judging GS will not understand the actual case and they will “stick it” to GS. And in any case that trial will be years down the road long after the damage is done. I mean who here remembers that Arthur Andersen was found not guilty of all charges in the end. Or that all those people convicted after the dot.com bubble were actually convicted of tampering with evidence not with insider trading or fraud.

    As to your final point, of course I would sue. Not because I have a case but because I would be trying to blackmail my vendor into making good some of my losses which sadly is a successful “investment” strategy in bubbles and one of the main reasons we keep having them.

  77. Danny Black writes:

    via Alea blog…

    This i think states the case pretty well. Note that GS incentives are not that far away from those of ACA and IKB.

  78. Danny Black:
    The SEC brought this insignificant case to help Goldman push the Dodd bill through Congress. If the SEC was serious about protecting Joe Schmoe, it would be screaming at the SDNY US Attorney to indict Goldman, AIG, Henry Paulson and Tim Geithner over the AIG fiasco. That’s big dollars. $180 billion! That’s where Goldman took the public to the cleaners. This case is Goldman’s not wanting to be “thrown in the briar patch” as Junior at Junior Deputy noted. The SEC knowingly and wilfully chose a weak case. Why? The public wants some blood spilled, so the SEC gives it a ketchup stained shirt.

  79. Philip writes:


    Thanks for the link to the — Powerpoint.

    I’ve pasted below the SEC’s own description of its case against Goldman. I think it’s pretty darn clear that the SEC case does not conform with your characterization that the case is centered on Goldman’s failure to disclose Paulson’s OPINION about the ABACUS deal.

    Instead, the SEC has focused on the allegation that Goldman failed to disclose that Paulson (a hidden party to the deal), betting ABACUS would collapse, had worked through what Goldman presented as an impartial third party (ACA) to hand pick mortgages Paulson thought were more likely to fail.

    Paulson’s OPINION is purely secondary to the charge. The SEC’s primary focus is that Goldman failed to disclose that Paulson was a party to the deal with an interest opposite to IKB’s and that he had an opportunity to tilt ABACUS in his favor by compromising ACA’s independence.

    Here’s the SEC quote:

    “The SEC alleges that Goldman Sachs structured and marketed a synthetic collateralized debt obligation (CDO) that hinged on the performance of subprime residential mortgage-backed securities (RMBS). Goldman Sachs failed to disclose to investors vital information about the CDO, in particular the role that a major hedge fund played in the portfolio selection process and the fact that the hedge fund had taken a short position against the CDO.

    “The product was new and complex but the deception and conflicts are old and simple,” said Robert Khuzami, Director of the Division of Enforcement. “Goldman wrongly permitted a client that was betting against the mortgage market to heavily influence which mortgage securities to include in an investment portfolio, while telling other investors that the securities were selected by an independent, objective third party.”

    “The SEC alleges that one of the world’s largest hedge funds, Paulson & Co., paid Goldman Sachs to structure a transaction in which Paulson & Co. could take short positions against mortgage securities chosen by Paulson & Co. based on a belief that the securities would experience credit events.

    According to the SEC’s complaint, filed in U.S. District Court for the Southern District of New York, the marketing materials for the CDO known as ABACUS 2007-AC1 (ABACUS) all represented that the RMBS portfolio underlying the CDO was selected by ACA Management LLC (ACA), a third party with expertise in analyzing credit risk in RMBS. The SEC alleges that undisclosed in the marketing materials and unbeknownst to investors, the Paulson & Co. hedge fund, which was poised to benefit if the RMBS defaulted, played a significant role in selecting which RMBS should make up the portfolio.

    The SEC’s complaint alleges that after participating in the portfolio selection, Paulson & Co. effectively shorted the RMBS portfolio it helped select by entering into credit default swaps (CDS) with Goldman Sachs to buy protection on specific layers of the ABACUS capital structure. Given that financial short interest, Paulson & Co. had an economic incentive to select RMBS that it expected to experience credit events in the near future. Goldman Sachs did not disclose Paulson & Co.’s short position or its role in the collateral selection process in the term sheet, flip book, offering memorandum, or other marketing materials provided to investors.”

    I think what you’re smelling is not politics by the stinch of Goldman’s corruption.

  80. Danny Boy writes:

    “The point was to have a public whipping boy whilst the Democrats ram a half-arsed regulation bill through.”

    On the politics of it, I couldn’t agree more. It’s pure kabuki. Obama apologists will say the the SEC is going to try to flip the VP and go up the food chain to the executives. We’ll see how that works out.

  81. Philip writes:


    I meant to address the Duffie presentation you provided the link to:

    * Duffie says that “Goldman took the risk that it would have some “leftover” long
    position, which turned out to be RMBS portfolio losses between 45% and 50% of the total pool size.” This is incorrect. The risk Goldman took represented about 5% of the total pool size.

    * Duffie says “Diversification across the portfolio of 90 RMBS also helps meet [Paulson’s] investment objective, because “noise” surrounding the idiosyncratic performance of individual pools would then be “averaged out,” leaving mainly the essential macroeconomic performance of the
    sub-prime mortgage sector.” This would be true if it accurately described what Paulson did in selecting the mortgages included in ABACUS. It does not. Paulson did not attempt to “diversify” the portfolio but to narrow it to include as many of the highest risk mortgages he could identify. This fit his investment objective much better than diversification did.

    * Duffie says “Subject to diversification and a given CDS rate (his cost), Paulson would benefit from the selection of RMBS that he believed likely to have large default losses.” True enough, but Duffie ignores the more important point that Paulson would benefit from the selection of RMBS that he believed to have a high likelihood of defaulting. Big default losses were good; high odds of default were better.

    * Re: ACA and IKB’s incentives, Duffie says “The structure of the ABACUS 2007 AC-1 RMBS portfolio would meet ACA’s investment objective. Absent a general sub-prime meltdown, the pool of mortgages underlying a given RMBS would not likely suffer large losses, in which case the mezzanine tranche would not be severely affected.” This would have been true if Paulson had not compromised ACA Management’s independence in its selection of RMBS to include in ABACUS or if Goldman had disclosed that Paulson had done so.

    * Duffie perpetuates the confusion between ACA Management and ACA Capital. They are different companies with different incentives.

    * Duffie says “Fixing its fees and the CDS premia, Goldman had an incentive for the RMBS portfolio to be well diversified and likely to have at most moderate losses.” This statement is incorrect. Goldman had an incentive to make this deal happen. The client who brought the deal to Goldman, Paulson, wasn’t interested in a “diversified” portfolio but rather one with as many of the highest risk mortgages he could get Goldman and ACA Management to stuff in the portfolio. Goldman’s incentive was to meet Paulson’s minimum requirements to proceed with the deal. Furthermore, Goldman recognized that Paulson’s approach was an innvoation and held the prospect of more similar deals. So a stream of fee revenue, not a single fee, was at stake with ABACUS.

  82. Danny Black writes:

    Phillip thanks for the quote. I have to say wow, the SEC case is an even bigger pile of shit than I imagined.

    1) The product wasn’t that new… In fact it was a variation on a relatively common theme as Mr Khuzami should know given he was heavily involved in creating them at Deutsche.
    2) Paulson and Co were in no way shape or form “one of the world’s largest hedge fund” at the time of the deal. They are NOW because they made a correct bet. Note they were making EXACTLY the same sort of bets in 2006 and they lost money. I assume SEC will be suing on behalf of those deals to because as we know Paulson’s OPINION is really key. This a key example of trying to portray what we know NOW as inevitable 3 years ago.
    3) The three main investors were ACA – who we know knew about Paulson’s involvement and his goals because he told them in meetings with their portfolio analysts – GS themselves – who I think we can also assume knew but hoped to lay off the risk – and IKB – who I will wager money also knew but also didn’t care.
    4) ACA were clearly independent as they rejected nearly half Paulson’s SUGGESTIONS and clearly they didn’t rate his negative views on the other reference names that highly. So the idea that Paulson selected this portfolio is clearly false.

    Finally, what they are saying is exactly what I said that Paulson’s OPINION was material and that it was not disclosed. As I said all of the market practioners I know can’t fathom WHY Paulson’s OPINION would matter. At the time, he was one of 14,000 hedge funds and not a particularly successful one. He was quoted widely believing these particular investments would crater and as of the time of Abacus closing he was WRONG, ie he LOST money doing these deals. ACA and IKB were INVESTING in these products which means they believed the products would go up in value and so presumably would disagree with the articles Paulson was writing at the time saying they would go down.

  83. Philip writes:


    I’m surprised to hear that you had not yet read either an SEC summary of its case or the filing documents themselves considering the authority and confidence with which you have characterized their case.

    ABACUS was in fact an innovation that, simultaneously, a few others beside Paulson were shopping around Wall Street trying to find underwriters to accept and having considerable difficulty doing so. Paulson, as a matter of fact, was turned down by two other IBs (on ethics and reputational objections) before GS accepted the deal. I recommend reading The Big Short and The Greatest Trade Ever for this background.

    Your point #3 has been disposed of previously by Steve in his excellent discussion of what represents formal disclosure.

    The rest of your points are immaterial to the merits of the SEC case.

    Best wishes.

  84. Danny Black writes:

    I skipped the press release and just read the propectus and flipbook and the claim that somehow Paulson’s view was material. I have never read either of those books but then I work in a closely related field so not sure what i would have learnt.

    Couple of things seem to confuse you with Duffie. Less 45% losses, GS is covered. Ditto more than 50%. Say the losses are 47%, GS swallows 45-47% part of the losses.

    Also you misunderstand what diversification means in this case. Paulson wants to avoid bonds backed by the only 1,000 guys with ARMs who happen to win the lottery and keep making their payments when – as he expected – interest rates get reset. He views ARMs in these geographic locations to be LIKELY to default so he was a cross-section of those ARMs. They also need to default in a relatively timely manner because otherwise his fund is likely to get closed down – remembering this is 2007, Paulson is a nobody in the credit market and had until the time of this deal a track record of LOSING money on these bets. This is the “noise” Duffy is talking about.
    Similarly ACA doesn’t want bonds backed by the mortgages of the 1,000 guys who work at the same company that goes bankrupt.

    Not sure what your 3rd point.. Paulson needs the defaults to be 21% to 100% cumulative loss. So they need to not only default but also have a low recovery value. So they need to make a loss not just default.

    Again with the 4th point. You seem to have difficulty getting your head round this concept. ACA and IKB didnt believe there was going to be a general subprime meltdown. If they did they then they would not have invested in these bonds. Again, I know it is hard but in 2007 when this deal closed these two firms were the “geniuses” in the market who had made outsize returns for their shareholders. Paulson was the “idiot” for believing what the market didn’t. Paulson was the “idiot” for going into a market he knew nothing about and pig-headedly continuing to lose his customer’s money. It is only LATER we discovered he was right and others were wrong. In the market, different people have different opinions. One of the 14,000 hedge funds was bound to bet the right way and they would have to use one of the many IBs and it would probably be through a vehicle like this. And bingo for the SEC and democrat agenda it happens to be a US bank who didn’t lose too much money overall and who we all hate anyway. Imagine if it had been Citi? Or for instance Khuzami’s alma mater who actually DID make a lot of money betting this way – hence why Mr Khuzami KNOWS that the statement about this product being “new and innovative” is a complete lie.

    To reiterate the above point, what Paulson thought was irrelevent. As investors, ACA and IKB should be be judging the investments on their merits. They had the skills and qualifications to do so – hence why GS keeps saying they are sophisticated. They were bonds that in ACA and IKB’s view were not likely to lose more than 50%. If ACA and IKB thought they would then they would not have bought them full stop. The information they needed to make a decision was all there. Paulson had no secret information they didn’t, it just they had a different view from him. Is the SEC really saying that if ACA and IKB had known – and again I will bet they did – that they would have thrown their hands up and said well if Paulson with his one year track-record of abject failure in this field thinks these bonds are bad then who are we with many years of success to disagree….

    Paulson’s approach was NOT an innovation. Other banks were doing it, other funds were doing it. Synthetics had been around for a while – in this industry a while is more than 12 months. If you are under the impression that Paulson would be so grateful to GS he would give his future business to them then you are sadly deluded about how transactional banking works. GS will boast about how “innovative” the deal is – which they would say about pretty much anything – but the focus on this deal is the fee vs the risk on the slice they had to keep to get the deal done. I will bet that there was internal discussion about that risk – which weirdly the SEC didn’t release – and that they considered it a low risk vs the fees.

    Again, what the SEC are relying on is that you will view these events through the googles of knowing the outcome coupled with releasing some embarassing emails which i can guarantee you’ll find in every single investment bank in the world.

  85. Danny Black writes:

    Phillip, actually I thought this was one the weaker posts here. It is IKB and ACA who have a responsibility to their shareholders, past present and future not GS. If IKB and ACA didn’t do proper analysis then THEY are the ones in breach of their fiduciary trust not GS. I mean when you invest in a mutual fund do you expect to get told about every deal they are thinking about investing in? You either believe the fund manager is competent or you sell your units/shares.

  86. Bet365 writes:

    Nicely written and thanks for this interesting post. Do you compose posts yourself or do you outsource?

  87. Fantastic report,I just now subscribed to your feed.

  88. Philip writes:


    Taking some of your points one at a time:

    * “Couple of things seem to confuse you with Duffie. Less 45% losses, GS is covered. Ditto more than 50%. Say the losses are 47%, GS swallows 45-47% part of the losses.”

    No, I him fine. His graphical presentation is consistent with your understanding, but his prose description is not. That was my point.

    * “Also you misunderstand what diversification means in this case. Paulson wants to avoid bonds backed by the only 1,000 guys with ARMs who happen to win the lottery and keep making their payments when – as he expected – interest rates get reset.”

    If I’m ever charged with a crime, I hope you’re on my jury. I don’t see how this statement can be anything other than speculation unless you have some special inside knowledge of how Paulson made his choices. I certainly don’t, but then the SEC does have special inside information, and they’re saying Paulson loaded Abacus with mortgages far more likely to default, not that he made selections designed to avoid mortgages that were far less likely to default. Big difference.

    * “Similarly ACA doesn’t want bonds backed by the mortgages of the 1,000 guys who work at the same company that goes bankrupt.”

    The way these mortgages were sliced and diced before Abacus was created made it impossible for “1,000 guys who work at the same company” to be in the same reference group. BTW, do you know how many mortgages were encompassed in Abacus?

    I’m curious how you’re so confident that you understand Professor Duffie’s perspective. Were you at his presentation or are you a student of his?

    * “Again with the 4th point. You seem to have difficulty getting your head round this concept. ACA and IKB didnt believe there was going to be a general subprime meltdown. If they did they then they would not have invested in these bonds.”

    These points are so elementary to this discussion, I’d have to be brain-dead not to understand them. Those who happen to disagree with you may not be as dim-witted as you think, though I do understand why one might wish to think so.

    * “Again, I know it is hard but in 2007 when this deal closed these two firms were the “geniuses” in the market who had made outsize returns for their shareholders. Paulson was the “idiot” for believing what the market didn’t. Paulson was the “idiot” for going into a market he knew nothing about and pig-headedly continuing to lose his customer’s money. It is only LATER we discovered he was right and others were wrong. In the market, different people have different opinions.”

    OK, I agree. So what. None of this bears on the allegations against GS contained in the SEC suit.

    * “And bingo for the SEC and democrat agenda it happens to be a US bank who didn’t lose too much money overall and who we all hate anyway.”

    Fact is that Goldman was very generous with the Dems, both with campaign contributions and jobs after government service. And I know quite a few GS alums, worked with them, and respect them. I’m not a GS hater, but I think they’re guilt as charged. Not because I’m biased or have 20/20 hindsight, but because I think I have a pretty good grasp on what was happening in the industry at the time and I’ve done some homework on the facts of the SEC case.

    Are you equally unbiased in terms of Goldman, Paulson, the SEC and the question of the legitimate role of government in regulating the banking industry? Or are you an advocate of laissez-faire policy? Or have a financial stake in one of the parties?

    I think that’s enough for now.


  89. […] guessing no one is thinking, or cares to think, about Goldman Sachs these days.  However, this article helped me understand the deal in question better than anything I had read […]

  90. David Larsson writes:

    “I’m not sure about who’d win an unpopularity contest, but the lawyers seem to have one up on bankers for ethics, at least in theory…”

    I don’t mean to hold lawyers out as ethical role models. I remember the moment in my last year of law school when I found myself sitting in room full of people trying to reach consensus on an issue: none of us were listening to each other; each of us was trying to “one-up” the last one; all in all, it was a very unpleasant experience. I remember thinking to myself: “Oh, no! I’m in a room full of … lawyers!!! On, no! I’m going to be in this room the rest of my life!!! What have I done?!?”

    When I went to law school, I visualized “Legal Ethics” as a course in which we would all wear togas and think deep thoughts about important subjects; it turned out to be more along the lines of “How Can I Take Possession of the Murder Weapon and keep my license to practice law?” But over the years (I, too, am middle aged, but only if I live to be 112), I have come to appreciate the rules of professional responsibility because they provide tools with which one can at least begin to address the many thorny problems — business, legal, ethical, and, yes, even moral — that inevitably emerge when one tries to find “the right thing” to do in any given situation for one’s client, for one’s law firm, for oneself — even for one’s adversary.

    I haven’t yet seen much evidence of the existence of such tools for investment bankers and/or investment advisers in the GS / Abacus controversy. I was surprised at the shallowness of the discourse during the first hour or so of the GS hearing about GS’s “duties” to its “clients.” Sens. Levin and Collins are intelligent people, and so were folks in the first GS panel. I would have expected the discussion to reflect rich, even nuanced, reasoning about the applicable rules and guidelines for identifying and carrying out those duties, but I did not hear much of that. I find that both curious and unfortunate, because the situation is, in some ways, not all that different: as Steve has observed, unless those bankers / advisers are, say, George Soros playing strictly with his own money, almost all are acting as agents on behalf of one or more principals: whether or not the identity of the principals is disclosed to the other party.

  91. Philip writes:

    David Larsson-

    I thoughtful post. I think IBs (and many other American industries, for that matter) have adopted the post-modern ethical credo of “the sole responsibilty of management is to create shareholder value.” Nothing about customers, nothing about employees, nothing about the communities in which they exist. And the guys who follow this credo are the “good guys” these days when viewed in the context of the legions of companies whose shareholders have been wiped out by management’s recklessness while management received enormous compensation packages and walked away from the wreckage without a scratch.

    Concepts like “fiduciary”, “trust”, “integrity”, “client-centric” are hopelessly old-fashioned, except to the degree that they’re useful in sales and public relations.

  92. David Larsson writes:

    Thanks, Philip. I’m sure you’re correct. Funny that, as you suggest, the post-modern ethical credo seems to be working much better for individual senior managers (“enormous compensation packages”) than it does for shareholders. So why do shareholders continue to pay the enormous compensation packages? Wouldn’t Adam Smith’s Invisible Hand of the Marketplace kick in, and we’d see a shift toward management by an infinite number of modestly compensated monkeys?

    It makes me think of Alan Greenspan’s October 2008 Congressional testimony: “Those of us who have looked to the self-interest of lending institutions to protect shareholder’s equity (myself especially) are in a state of shocked disbelief.”

  93. Danny Black writes:

    David Larsson, maybe you misunderstood the purpose of the hearing. Levin et al have zero interest in understanding GS’s business, rather they had an interest in actively misrepresenting it. Everyone with more than two brain cells knows this.

    GS is basically a collection of companies:

    1) Asset Manager, they have a fiduciary responsibility to the investors in their funds.
    2) Primary issuance advisor have one to their client who is issuing the product
    3) Market maker, whom market professionals are approaching and who is approaching other market professionals. These people have an obligation to provide a two way price. They have an obligation to the shareholders of GS not to lose money. They have an obligation to accurately describe the product so that the counterparty knows what he/she is buying. They are not advising the client on whether they should do the trade or not – they may be trying to convince the client to do it but the client should realise that this is not “investment advice”. This arm is committing fraud if they say it is an IBM call option when in fact it is an IBM put option. It is NOT committing fraud if it sells you IBM stock but internally thinks the price is going down.
    4) Some corporate finance bits which would have a duty to their clients because it is providing advice.

    The are really two points here and i clearly have not been able to describe them well and apparently if you don’t work in the markets they are impossible to grasp and i have no idea why as i think they are pretty obvious:

    1) There is simply NO such thing as a product “designed to fail”. Doesn’t exist. You can simply have a view or opinion about the future and structure a product to maximise the payout if that opinion is correct. But your opinion might or might not be correct. Doesn’t even matter if your view is very likely to be correct. If a bond has a 99% probability of being worth zero and 1% chance of being worth 100 then – ignoring opportunity cost – I should be happy to offer you 0.1 cents for it. As long as I know what I am buying – and IKB and ACA did – then as an investor thats all I need to know.
    2) Things that are obvious now are not obvious in the past. It is hard when you know the outcome not to have the past coloured by that knowledge. It is like seeing a film when you know the end, suddenly things that were irrelevent the first time you see are important in retrospect. At the time of this deal Paulson was a nobody in the credit markets. He was a man with a track record of failure in credit. A man clearly out of his depth. There is no reason to believe that were we to transport ourselves back to 2007 when this deal was happening that you would give weight to his ideas. We NOW know he was right, we NOW know that he would make billions on these bets and he is NOW one of the largest hedge funds. The SEC is hoping you will take that post-hoc knowledge and let it colour your view of the transaction in 2007.

    Let me give you an example of the second point. I did a degree with a large component of history. In one of the courses, the lecturer was throwing a minor hissy fit about a transaction in 1905 where one guy paid the shah of Iran for the right to all the oil in Iran. The Iranians were clearly being screwed. After all we know that that Iran has massive oil deposits, we now know that investment would end up 50% owned by the UK goverment and become BP – one of the largest oil companies in the world – we now know oil is one of the vital lubricants for the world economy. So clearly a slam dunk that the shah was screwed in the deal right. Except:

    1) Nobody at the time knew there was oil in Iran
    2) In fact the reason the UK goverment ended up owning half the investment was because the original owner went bankrupt whilst he paid the fees to the shah whilst failing to find any oil. The Treasury official responsible for the investment – a then relatively unknown Winston Churchill – nearly faced political oblivion over the decision
    3) In 1905, the major use for oil was heating. It was a declining market with electricity coming online. There were no oil based ships, planes hadn’t been invented and the internal combustion engine was only just becoming viable.
    4) And finally even if the above wasn’t true the shah had no way to get the oil out of the ground.

    So for the shah that was the slam dunk bet of the century. This had to be the closest thing to designing a one sided deal there is and the shah lost.

    Philip, you clearly can’t understand the above but given that you believe that the future can be predicted with 100% certainty and that outcomes one year later are material to transactions now you should start your own hedge fund shorting these “designed to fail” products. Someone with your genius would be a billionaire in no time. I would be happy to be your counterparty anytime of the day.

  94. Danny Black writes:

    Phillip, I wonder what industry you work in that is so highly ethical.

    I do know Duffie, attended many of his talks and he is a model of crystal clarity as the presentation posted is an exemplar of that. I thought it is everything he said is obvious to anyone with two or more brain cells to rub together. I am frankly amazed you seem not to be able to get it.

    If Paulson’s reputation and IKB and ACA view of him is so irrelevent to the case then why could you possibly think that his view of the performance of these bonds IS material?

    My bias is I work in IB, in derivative field. Looking at the bonus figures at my colleagues there, sadly not at GS.

    One point I should make is that the idea banks hate regulation is utterly false. We LOVE regulation. Especially poorly thought out, knee-jerk regulation. A significant portion of my business would not exist without it.

  95. csissoko writes:

    Danny Black:

    Enough already. It’s abundantly clear that in it’s relationship with IKB, GS was acting as a placement agent of the newly issued Abacus 2007 AC1 security created by Goldman. Marketing making and issuing securities are distinct activities with distinct responsibilities.

    Not only that as zerobeta has pointed out, there’s no evidence what-so-ever that Goldman was doing any market-making in the Abacus transaction. Was IKB told what Goldman’s bid and offer were for the security — or at least what Goldman’s spread was between the bid and offer?

    Just because Goldman wants to confuse Congress about what a market maker does, doesn’t mean the rest of us need to thoughtlessly believe their nonsense.

  96. Danny Black writes:

    csissoko, read Coffee’s comments before. They didn’t come to GS for its skill in CDOs because it was not picking the underlying reference portfolio, ACA selected the underlying portfolio and they were the ones who were meant to be the ones with the “expertise and skill”. They explicitly were NOT coming to GS for advice on whether this was a good bet or not. They also have no duty to tell people that the other side is short – although this is obvious from the construction of the synthetic CDO.

    Also he should be aware as you should be that GS in an underwriting role has a duty to accurately describe the product being sold – which they did. Care to point out where the product did not behave exactly as described?

    Feel free to thoughtlessly believe whatever you like.

  97. csissoko writes:

    DB: I never claimed that IKB came to GS for advice, so your first paragraph is irrelevant.

    “GS in an underwriting role has a duty to accurately describe the product being sold – which they did” Say, what?! Neither you, nor I, nor any lawyer can make that claim as if it were a fact. The place where GS underwriting obligations are finally determined is in a court of law and while it is possible that your claim will be substantiated, we all know that courts and judges are complicated places and since the underwriters obligation in a synthetic ABS CDO has, I believe, never before been considered — we are all of us on unknown territory here.

    BTW your comment led me to actually post a rant I’d been sitting on for a while: http://syntheticassets.wordpress.com/2010/05/11/the-myth-of-the-market-maker-in-cdos/

  98. Danny Black writes:

    csissoko, exactly what bit of the product did the investor not know about? Not talking about the motive of one side of the trade – which is not relevent – but the ACTUAL INVESTMENT they were making. In what way did it not behave as described? What reference obligation did GS fail to disclose in the document? What trigger was there or wasn’t there that wasn’t written up? etc etc etc. They had ALL the information they needed to make an informed investment choice and they chose wrong – in hindsight.

    Lots of products are thinly traded – certain corporate bonds, certain gilts, longer maturity IRS – and alot of the time you don’t trade them you trade something that gives you the same economic result – modulo counterparty and model risk and usually some basis risk. In the case of CDOs you’d normally trade a CDS on it or on a related index. In the case of this one you’d probably trade the CDSes underlying it. In the prospectus it does say that GS undertakes to initially make markets but that it cannot guarantee it will in the future, that the market is likely to be illiquid and that the buyer should be willing to hold to maturity. So much for zerobeta’s claim.

    By the way, via DougChia – i think i spelt his name right – there is this analysis of the case. My own sample biased impression of people’s view of the case is the same as this guy’s and I think he is spot on with what the SEC will look to do and what GS would hope to do:


    As for marks, I can tell you in exotics investors always question more anything that doesn’t give them a profit or at least break even. 99 times out of 100, offering to sell more of the same at THEIR marks usually shuts them up.

  99. csissoko writes:

    DB: Your first paragraph is a very good summary of the defense lawyer’s brief — and just as one-sided.

    The problem with so-called market making in CDOs is that the investment banks didn’t carry any trading inventory of second-hand CDOs. They carry corporate bond trading inventories and Treasuries of every maturity. Why? Because they are genuinely market making in bonds. Obviously we can pick out a specific bond issue and find that they have no inventory, but that is completely irrelevant.

    Claiming that market making in CDS is equivalent to market making in CDOs is interesting: Have any IBs actually made that claim? I suspect this would put them on very, very dangerous legal ground.

    Thanks for the link. Grundfest is clearly another professor who needs to read SRW, here. (i) He conflates ACA capital with ACA management and (ii) by doing so, he makes his only defense of the IKB case completely irrelevant. Also, he claims (slide 11) that IKB contributed to ACA’s selection of the underlying — I suspect that this is just wrong.