Felix Salmon

How Steve Cohen moves stock

By Felix Salmon
November 26, 2012

Eric Hunsader, at Nanex, has managed to put together some fantastic charts of what exactly happened in Elan, the stock at the center of the latest big insider-trading case.

First, here’s the big picture:


The red arrow shows the period “throughout 2007 and up to July 2008″ during which SAC “established a substantial long position” in Elan. The blue arrow points to the frantic week during which SAC sold off more than its entire holding, ending up with a significant short position, just before the stock plunged.

When Elan opened for trade on Monday July 21, 2008, it was at a multi-year high of more than $35 per share, and SAC’s long position was massively in the money — after all, they had been buying since it was less than $15. And then SAC started selling, aggressively.

Over a four-day period, SAC sold its entire position of 10.5 million shares between Monday and Thursday, at a super-high average price of $34.21 per share. The head trader, who said that he sold the stock “quietly and efficiently through algos and darkpools”, continued to sell. By the end of the trading session on the 29th, he had sold more than 15 million shares for more than $500 million. The complaint notes that the SAC trading “constituted over 20% of the reported trading volume in the seven days prior to the July 29 Announcement.”

What does that kind of massively one-sided selling do to a stock price? This:


Basically, Elan moved sideways for most of the time that the stock was being sold. Day 1 was great, Day 2 was decent until the end of the day, Day 3 started off well but then deteriorated, Day 4 was horrible, Day 5 was much better, Day 6 had a good morning and a gruesome afternoon, and Day 7 was pretty good.

And by the end, in the wake of $500 million of concerted selling in a pretty illiquid stock, the share price was about $33.50 — pretty much exactly where it was on the Friday before the selling started.

Eric’s detailed day-by-day charts are well worth looking at, but for me there are two big-picture lessons here. The first is that SAC is an amazingly good trading shop; we probably already knew that. And the second is that any time you see a market reporter blaming “selling” for the fact that a stock went down, you can take that with a pinch of salt. Because the lesson here is that an absolutely enormous amount of very real selling can have a surprisingly small effect on a stock’s price.

8 comments so far | RSS Comments RSS

could be great trader, but sufficient facts are not in evidence… how did that price compare to the VWAP, and how did the stock perform relative to industry and market as a whole in that time period.

maybe some real traders can weigh in, but I don’t think you necessarily have to be a genius to trade along with VWAP at 20% of volume for a few days, hit a few bids in the dark pools if they are there.

when the market is strong and it lets you out, it lets you out… when it doesn’t, it doesn’t.

Posted by streeteye | Report as abusive

Agreed with streeteye. 20% participation is (often) not a challenging level to maintain with only minimal market impact. SAC may well have a killer desk, but this trade doesn’t necessarily prove that. Then again, I’m not at all familiar with Elan ADV over this stretch or what VWAP was or what price SAC got for their position… so what I’m saying is – lock him up, and I don’t care about his trading skill.

Posted by clearance42 | Report as abusive

“any time you see a market reporter blaming “selling” for the fact that a stock went down, you can take that with a pinch of salt. Because the lesson here is that an absolutely enormous amount of very real selling can have a surprisingly small effect on a stock’s price.”

That basically says supply and demand are irrelevant. The price was climbing when the sales were made, and it is likely the prices would have been higher if SAC’s shares weren’t on the market. The stock was at a multi-year high when they started to sell, which means people were optimistic about the stock, and given the positive feedback loop that is the stock market, there would be reason for the price to continue to climb. Unless there were more sellers than buyers.

Posted by KenG_CA | Report as abusive

At any given time 50% of the volume will be selling. So SAC was 40% of the selling. That’s substantive but they clearly had a bit of patience, selling over seven days in a world of high speed trading is hardly aggressive. As for algos all you really need is one that doesn’t cross the spread (eg something a skilled 8 year old could code) and doesn’t show too much at any given time (bit trickier, but not too hard if you have patience). Basically though its not hard to sell when there are buyers, the real art here is that SAC knew to buy at 15 and sell at 35, and that’s where the insider trading potentially comes in… If you are the only one with the bad news it’s not too hard to find buyers. It’s the people who get out fast when everyone knows the bad news that are good traders…

Posted by AbeB | Report as abusive

50% of the volume is selling? there’s a buyer and a seller on every trade, no? (unless the volume figures count that as 2 trades, a buy and a sell?)

Posted by streeteye | Report as abusive

Well said AbeB.

Posted by QCIC | Report as abusive

Yeah except he is wrong (AbeB).
100% of the volume is selling – otherwise what the the people who are buying, buying?
A school for ants?

20% of the volume is 20% of the volume.
Every trade is obviously a buy and a sell.
If 100 shares change hands the volume is 100 not 50.
If you are 20% of the volume you have sold 20 shares not 40.

Also the buying at 15 isn’t in question.
It’s the selling at 35 that matters.

However Abe is right that this data doesn’t show anything about how good SAC’s trading desk is.

Our rule of thumb is to stay under a third of the volume and you won’t move the stock.

Posted by TinyTim1 | Report as abusive

Too many of you pay attention to the wrong things. Too many wannabe economists mixed with egos.

Posted by SenorAlpha | Report as abusive

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/