Andrew Ross Sorkin has got his hands on what looks for all the world like a pro-forma piece of legal CYA out of Goldman Sachs, and is trying to turn it into a story:
For years, Wall Street whispered that Goldman Sachs profited handsomely by trading ahead of — or even against — its own clients.
On Tuesday, a Goldman executive made an unusual admission that, in some cases, the rumors were true.
In an e-mail message to select clients, Thomas C. Mazarakis, the head of Goldman’s fundamental strategies group, acknowledged that his unit often provided investment ideas that the firm had already traded on. Sometimes Goldman has even taken the opposite approach, betting against particular instruments that the group has recommended.
The good news is that Sorkin is enough of a blogger to actually publish the email in question. But weirdly I can see nothing at all in the email saying that Goldman ever trades against its own clients, or bets against instruments it’s recommending. Here’s the relevant bit:
We may trade, and may have existing positions, based on Trading Ideas before we have discussed those Trading Ideas with you. We may continue to act on Trading Ideas, and may trade out of any position, based on Trading Ideas, at any time after we have discussed them with you.
Now it’s conceivably possible that when Goldman makes a trade “based on” one of its Ideas it actually takes the opposite position to that implied by the Idea. But there’s nothing in the email to indicate that ever happens, or has ever happened — which means that it’s very hard to see how Sorkin can characterize the email as “an unusual admission” that sometimes Goldman bets against instruments it’s recommending.
I suspect that Sorkin might be looking at the “may trade out of any position” language — but of course when Goldman trades into a position it will necessarily trade out of that position at some point. Doing so will involve selling what was previously bought, or buying what was previously sold, so yes I suppose that if it’s selling while its clients are buying, then you could say it’s trading against its clients. On the other hand, any time a Goldman client buys a position from Goldman, or sells one to them, Goldman is by definition trading against that client. And that’s exactly what the client wants!
Alternatively, there’s a Clintonian reading of Sorkin’s story, in which it’s literally true but actually doesn’t mean what you think it means. The “unusual admission”, on this reading, is just that Goldman sometimes trades ahead of its clients, not that it ever trades against them. And the bit of the story saying that sometimes Goldman bets against instruments it’s recommended comes not from the Goldman email but rather from previous NYT reporting.
Either way, I think there’s a lot less to this story than meets the eye. The email is clearly legal boilerplate, and not an “unusual admission” of anything at all. Goldman will continue to share trading ideas with its biggest clients, and those clients will judge those ideas on their own merits, rather than entering into them just because they come from Goldman. Sometimes, the clients will decide to trade against the ideas. They’re all big boys, and they make their own decisions.
The idea that Goldman is conflicted because it trades with and against clients is ridiculous — that’s its job. It’s a broker-dealer, not some kind of fiduciary. I do think that Goldman is too big, but I don’t believe it’s remotely constructive to start attacking the fundamental role of any broker-dealer, which is to make markets by trading in and out of lots of positions with lots of clients. That’s one of the few ways in which Goldman actually makes the world a better place.