A Single Fake AP Tweet Can Apparently Crash the US Stock Market

Here's a fun thing we learned today: a single tweet can cause the stock market to begin to crash in about two minutes. Also, correcting that tweet will help it to bounce it back immediately.

Just after 1 pm, the Associated Press's Twitter account was hacked. A single faux tweet spewed out from the handle:

"Breaking: Two Explosions in the White House and Barack Obama is injured."

A group called the Syrian Electronic Army quickly took responsibility. Despite sounding nothing at all like a news wire headline, the message was, obviously, instantly retweeted thousands of times. Then this happened:

The Dow Jones Industrial Average dropped 150 points in a matter of minutes, erasing around $150 billion of value. Of course, once someone looked up at the White House and saw that it had not in fact exploded, and the press secretary working there had disavowed the report, the market corrected itself and bounced back to precisely where it was when it started.

One trader who watched the Dow drop tells me that this is exactly how markets are supposed to behave—they reacted quickly to the news, and recovered just as quickly.

That's true: markets should by all rights be thrown into panic if the White House were bombed. But this precipitous drop was spurred by a single, link-less tweet blaring unconfirmed (and dubiously worded) news. Even if it corrected itself just as quickly, the drop shows that it may be relatively easy to manipulate markets by hijacking just a single major wire agency's social media platforms. And it reiterates what we already knew—markets are hypersensitive and often moving faster than facts.

And that's because the market is now controlled by sophisticated robots. Computer programs assisting in high frequency trading operations were what reacted so quickly to the drop, not real human beings—most actual traders didn't buy the "news" for a second.

Jonathan Corpina, a senior managing partner with Meridian Equity Partners, told Bloomberg that he got on the phone with someone in D.C. to check the story out for himself:

“He did not know what I was talking about," Corpina said. "He said I’m staring at the White House and there’s nothing going on here right now.”

Algorithmic trading programs that read news headlines may have started the selling, he said. “And then other algos jump in to play the snowball effect, and little by little you have the computer trading systems that have canceled all their orders on the buy side and the sell algos hit all these bids, and that’s the big dip we saw,” he said.

Herein lies the pitfalls of high frequency trading, ladies and gentlemen. Bloomberg spoke to other traders who found it "quite scary" that the automated response to a tweeted-out rumor can result in such a sizable crash. Some warned against the use of high-frequency automated trading.

This incident looks to be an isolated stunt, and AP writers also report getting their emails and personal accounts phished. Still, the brief debacle demonstrates just how easy it is to screw with the increasingly automated programs lording over already fickle markets; and how a more concerted effort could feasibly push them even further.