How did Knight lose $440 million?
Well, that didn't take long. Knight Capital announced a pre-tax loss of
Wow: $440 million? How on earth did it lose that much? It's easy to see. I noted many times yesterday that the volume in the first half hour of trading on the New York Stock Exchange floor was titantic: shortly after 10 a.m. ET, it was about 300 million shares, well more than twice normal volume at that time.
This would indicate that the trading program that came out of Knight resulted in millions of duplicate trades. And it was not going to get much relief in the form of busted (erroneous) trades. The NYSE last night busted trading in only a half-dozen stocks and declared that "no additional symbols will be considered for review."
The NYSE was not being mean-spirited: The definition of what constitutes an "erroneous" trade is very carefully defined, and the NYSE determined that based on that definition only a half-dozen stocks fell into that category.
Once that happened, Knight was faced with tens of millions (at least) of trades that were not clearly "erroneous," but which it is likely liable to cover. At that point, it was just a matter of doing the math.
That's why it was able to come up with a number so quickly.
While losses may be settled for the moment, the longer-term risk is reputational. Knight has a very large retail customer order base which may begin to question its ability to execute.
Patrick O'Shaugnessy at Raymond James downgraded Knight from “strong buy” to “market perform,” saying "these events will limit Knight's earnings multiple expansion due to perceived risks of its market-making business model."
1) What exactly caused this "technology issue"? Knight still hasn't told us. I mentioned several times yesterday that the NYSE had launched a new program Wednesday, the Retail Liquidity Program (RLP), essentially a dark pool inside the NYSE. It is quite possible that Knight had a problem interacting with that program due to some kind of coding issue. Other firms apparently interacted with the RLP without incident.
2) Why did it take so long to shut the damn errant program off? It ran for almost 30 minutes! Traders on the floor noticed a problem within the first couple minutes of the open — I was standing there — but it appears to have kept running to almost 10 a.m. ET.
Huh? Traders seemed baffled that it kept running, but when I asked a couple guys, "Why don't they just call up the source of the trade and tell them to hit the kill switch?" nobody had an answer.
I think the problem was the NYSE DID contact Knight, and it may have taken Knight some time to identify the problem. If this is correct — and I think it is — this will give more ammunition to those who are arguing for more robust oversight over the trading systems ... at the very least better interactivity and more redundancy.
—By CNBC’s Bob Pisani
Bookmark CNBC Data Pages:
Want updates whenever a Trader Talk blog is filed? Follow me on Twitter: twitter.com/BobPisani.
Questions? Comments? firstname.lastname@example.org