Nanex has published a very important post titled May 6’th 2010 Flash Crash Analysis; Continuing Developments; Sell Algo Trades:
We have obtained the Waddell & Reed (W&R) May 6, 2010 trade executions from the executing broker in the June 2010 eMini futures contract. There were 6,438 trades totalling 75,000 contracts. We matched them by time, price and size to the 147,577 trades (844,513 contracts) in the CME time and sales data between 14:32 and 14:52 (they matched exactly). One-second resolution charts of the W&R trades along with other eMini trades are shown below in various time frames.
The SEC report identified a Sell Algorithm selling 75,000 contracts as the cause of the flash crash. If the “Sell Algorithm” in the SEC report refers to the Waddell & Reed trades, then there is a problem. A big one. Looking at the trades in context with the other trades during that time, they do not appear to be significant. The W&R trades also do not occur near the ignition point (14:42:44.075) we identified earlier. Furthermore, the W&R trades are practically absent during the torrential sell-off that began at 14:44:20. The bulk of the W&R trades occurred after the market bottomed and was rocketing higher — a point in time that the SEC report tells us the market was out of liquidity. Finally, the data makes it clear that the algorithm does take price into consideration; you can see it stops selling if the price moves down over a short period of time.
Something is very wrong here.
There are some charts of great interest.
I quite agree that something is very wrong here. The SEC stated quite explicitly:
This large fundamental trader chose to execute this sell program via an automated execution algorithm (“Sell Algorithm”) that was programmed to feed orders into the June 2010 E-Mini market to target an execution rate set to 9% of the trading volume calculated over the previous minute, but without regard to price or time.
Let’s play lawyer. “Target” does not necessarily mean that the execution rate was achieved. It only means it was targetted, and it is possible that other considerations adjusted this target. “Without regard to price” is formally consistent with Nanex’s observations regarding the appearence of pauses in algo execution, since the pauses are triggered by the rate of price decline; it might be that “without regard to price” means simply that once the market price has flattened out it starts executing again, regardless of what level prices reached during the pause period.
But that’s just hair-splitting. I will be most disappointed if this SEC report turns out to have been an exercise in legalistic hairsplitting.
Now, it is not necessarily the case that the SEC report and conclusion is completely wrong, even given the execution pause; it is entirely possible that Waddell Reed soaked up all the available liquidity and the price decline triggered other things. In that case the SEC report is not wrong, per se, but it has missed the point of the affair which (assuming that the hypothesis is correct) is far different from the stated conclusions.
After all, if I put in a market order to sell 100,000 shares; execution of this order takes the market price down a buck; the decline in market price triggers stop loss orders of 200,000 shares which take the market price down another two bucks; and this decline in market price causes the momentum players to jump in with both feet while all potential buyers sit on their hands and the maret price declines another four bucks … what’s the cause of the seven dollar decline? Me, I would say “mainly bozo price-based technical traders”, but maybe the SEC would blame me for the whole thing.
The SEC needs to clarify this.
The story has been picked up by the New York Times:
The findings are based on the actual private trading data from the afternoon of May 6 given to Nanex by Waddell & Reed — presumably because Waddell & Reed wants the data to be made public to clear its name as the cause of the crash.
The data was verified for Nanex by Barclays Capital, at the request of Waddell & Reed.
Barclays Capital supplied the computer algorithm used by Waddell & Reed to make the sale.In a statement, Waddell & Reed said, “Following the recent release of the regulatory report on the ‘flash crash,’ many market observers have noted that the events of May 6 involve multiple issues that transcend the actions of any single market participant. We agree with those observations.”
A spokesman said, “After discussion, we granted permission to use the data, which was supplied to us by the executing broker.”