The Wall Street Journal has a great scoop on A1 today reporting on a heretofore-unknown practice at Goldman Sachs called “trading huddles,” where it gives tips to big trading customers—tips that its research customers rarely hear about.
The paper leads with Goldman’s recommendations regarding Janus Capital Group, which a Goldman analyst rated “neutral” one day, told traders over the next couple of days to buy, then finally upgraded to “buy” in a report to all his customers six days later.
Goldman Sachs Group Inc. research analyst Marc Irizarry’s published rating on mutual-fund manager Janus Capital Group Inc. was a lackluster “neutral” in early April 2008. But at an internal meeting that month, the analyst told dozens of Goldman’s traders the stock was likely to head higher, company documents show.
The next day, research-department employees at Goldman called about 50 favored clients of the big securities firm with the same tip, including hedge-fund companies Citadel Investment Group and SAC Capital Advisors, the documents indicate. Readers of Mr. Irizarry’s research didn’t find out he was bullish until his written report was issued six days later, after Janus shares had jumped 5.8%.
The sequence goes: Goldman analyst reiterates his “neutral” rating on Janus. The next day he tells Goldman traders to buy Janus. A day after that, he lets fifty of Goldman’s best trading clients in on the tip. Five days later, he tells research clients that Janus is now a “buy.” Meanwhile, the stock has already risen 5.6 percent.
Or so the Journal says. It sure doesn’t look like 5.6 percent from its chart, so I looked up Janus’s closing stock prices from those few days in April 2008. On April 2, Janus opened at $25.08 a share and closed at $25.38. On April 8, it opened at $25.93 and closed at $26.38. Using the open, that’s a 3.4 percent gain. Using the close, it’s 3.9 percent. Maybe the Journal used the precise intraday “ticks” to calculate this. If so, it should have said that and its chart should have reflected it.
I would also note that Janus stock jumped sharply—more than 9 percent—the day before Goldman told its own traders to buy. Looking back through the archives, I don’t see any news event that would account for such an unusual spike. The WSJ doesn’t mention this but should have. It also should have noted that Janus shares declined despite Goldman’s “buy” recommendation on April 8, 2008.
I don’t mean for these quibbles to take away from the story. They don’t. The paper has done some solid legwork here. It gets a hold of internal documents and reports that these trading huddles weren’t some one-off thing—it’s happened with hundreds of stocks.
The paper uses a second anecdote from the same April 2, 2008, trading huddle—this one about MetLife. Both anecdotes show Goldman telling a select group of insiders (“anywhere from six to 60 clients ) to buy something days before officially disseminating a “buy” rating to its broader research-customer list. Stocks often jump on an important analyst’s “buy” recommendations. This would seem to raise serious questions, then, about insider trading.
Look at how Goldman plays it cute:
Compliance officers sit in on almost all the meetings, Goldman says. Research analysts say they have been guided on what language to use in the huddles. Words like “buy” and “sell” are to be avoided, while “run up,” “give back” and “oversold” are encouraged. Internal documents reviewed by the Journal initially tracked the trading-huddle tips as “buy” or “sell,” but now refer to them as “up” or “down.”
While we’re (supposedly) overhauling the financial-regulatory system, it seems to me that policymakers need to look—again—at the conflicts that riddle investment banks. Right now, we depend on something close to Scout’s honor for them to stay on either side of the so-called “Chinese wall.” Every time I hear about a Chinese wall I reach for my wallet.