Friday, April 16, 2010

Wow, this Goldman thing is nuts...

I've defended Goldman in the past, saying that you can't get mad at them for just being smarter than everyone else.

I still believe that to be true - in the areas they really were smarter than everyone else, which is a lot of them, they should be given credit, not disdain - but the revelations this morning (that they allowed John Paulson to custom-structure CDOs he wanted to short, and then turned around and sold the long side of the CDO to investors and didn't tell them that anything in particular was unique about the CDO) was pretty blatantly unethical. Fiduciary duty is pretty clear - if you're acting as someone's agent, you need to a) disclose all conflicts of interest and b) you have "an obligation to act in the best interest of another party ... whenever the relationship with the client involves a special trust, confidence, and reliance on the fiduciary to exercise discretion or expertise in acting for the client."

The question yet to be answered, of course, is exactly how Goldman was contracted by the purchaser of the CDOs (the side that WASN'T Paulson's short). Still, there's an implicit contract when you buy a CDO package from structurers that the bank will structure it appropriately and ensure the buyer is getting something representative of what they think they're getting. Even if GS' conflict of interest is minimal - everyone knows GS gets paid a commission, and that they'll either hold or sell the opposite side of the trade - the disclosure and requirement that they act in the best interest of their clients is a pretty big hurdle, and although we'll find out more, it doesn't look like they cleared it. I'm not condemining them yet, because it's possible they did appropriately disclose the particulars of the CDO up to the standard they had to, and that the buyers should have done their due diligence. But it's a valid question whether that happened, and it'll be interesting to see what comes of it.

For example, here:  it says "But the deck was stacked against the Abacus investors, the complaint contends, because the investment was filled with bonds chosen by Mr. Paulson as likely to default. Goldman told investors in Abacus marketing materials reviewed by The Times that the bonds would be chosen by an independent manager. "

Paulson was anything but independent, by the financial definition of independent, so that would in fact constitute a breach of fiduciary duty, as well as fraud, if it were true. Of course, the people going over the marketing materials may not have materials relevant to this particular Abacus CDO, just a different Abacus one. The materials are also being reviewed by journalists, not securities experts. Take it for what it's worth - suggestive but not bulletproof.

There are also rumors flying around that Paulson and Goldman did the same thing with Greek debt - Goldman structured it, Paulson took the short end and let Goldman sell the long end. Not enough information to say, yet, but this CDO thing does seem to be very telling. It doesn't sound like Paulson will get into legal trouble, because it was Goldman's responsibility to find and inform a counterparty. Still, the whole thing stinks...

We probably should have seen this coming. Both of these were dated in February 2010.

Another point - the idea that Goldman would sell or hold the opposite side of the trade is actually not a perfect argument either... a lot of these things were sold by people to hedge their risk, not to bet against the securities. Thus, there's a major implicit difference between "there's one seller betting against this group" and "there are individual sellers betting against each member of this group". 

Additionally, and this is something I definitely need to think through a little more because I don't know the details about this CDO: typically, sellers are only betting against CDO or synthetic CDO securities insofar as someone who sells me a share of IBM is betting against IBM by selling it to me - it's not a contract, it's a security. If it were a contract - a credit default swap, or even an option, you know someone's taking the opposite side, but with a security, it's less clear. There are lots of good reasons to sell a security (liquidity, etc), but many fewer to short the security (hedging, or betting it's going down). Thus, buyers shouldn't necessarily have to assume they're taking the long side of a short trade when they are plausibly (in fact, most of the time) just taking a security disposal.


  1. This may be more reflective of my lack of understanding of this side of finance than anything else, but did Goldman really owe a fiduciary duty to the sophisticated investors buying the long side of the CDOs?

    To what extent were they simply the people who had the asset that the others wanted to buy, and to what extent were they acting on behalf of the institutions buying the CDOs when it came to selection, and whatnot. If the former, and if they were transparent about the product and didn't make any misrepresentations about it, then I'd call the situation unseemly but not necessarily unethical.

    I'm legitimately curious, as this is something that I don't have much knowledge of.

  2. Okay, having educated myself further, I renounce the points tentatively made above.

  3. Just because a client should have been smarter doesn't mean the client should be lied to. It's like accounting rules - theoretically, anyone investing anything should at least read the accounting statements and question what could be wrong about them. However, an executive who lies about accounting is still a liar and deserves to face legal consequences.