Back in April of 2013, the Associated Press sent out a horrifying tweet: there had been two explosions at the White House and President Obama had been injured. Naturally, the tweet claimed a great deal of attention, with as many as 4,000 retweets in the span of a few short minutes. Now, if you’re reading this wondering how you completely missed that time in April that somebody tried to blow up the president, you’ll be relieved to know that the incident never occurred, a fact which the Associated Press itself picked upon very quickly as they soon removed the offending item and released a simple statement of apology. The tweet which had alarmed so many, especially considering its proximity to the recent Boston Marathon bombings, was nothing more than a prank perpetrated by a hacker. However, as far as pranks go, this one was considerably more impactful than most.

As stated above, it only took a couple of minutes for the Associated Press to notice and remove the falsified tweet. Within that minute span of time, this single tweet caused the Dow Jones industrial average to drop by around 150 points. That’s around $136 billion dollars, gone in the blink of an eye. That’s the kind of event that can cause a depression. Fortunately, the market rebounded a few minutes later when the report was revealed to be a hoax and all was as it had been before the insanity. We didn’t spiral into a financial meltdown, though that’s certainly where we’d been heading. And that’s the end of that particularly distressing story, until it happens all over again.

Sure, it’s scary to consider what might have been had things played out differently, but it’s even scarier to consider that it could very easily happen again. High frequency trading is big business. Those who participate in it are tasked with placing hugely important bets, the kind of bets which can have a huge impact on the entire world, within milliseconds. These traders make such a miniscule amount of profit on each trade that if they want to make any money at all they have to make thousands or even millions of trades a day. Keeping that in mind, it’s easy to see why they have computer programs to do much of that trading for them. These computers, working from a list of selected keywords like “bankruptcy” and “Barack Obama,” comb through hundreds of news sources, including some of the more “reputable” Twitter accounts out there, looking for data which might influence the market. It’s understandable that an explosion at the White House which injured the President could fall within those parameters.

After the dust settled and everyone made their apologies traders began to sift through the wreckage and attempt to figure out what had just happened. In the end, they came to the conclusion that perhaps they had been a bit too hasty to make Twitter such an integral part of the trading process. As Jack Ablin, the chief investment officer at BMO
 Private Bank, put it
, “False reports due to the account hacking undermine Twitter’s credibility in some respects. We have to recognize Twitter for what it is, a social media site and an unfiltered news source.” On one hand, this “revelation” seems completely absurd. These people didn’t realize that Twitter wasn’t a perpetually accurate information machine? You’re going to trust the people of this site, where a few months later hundreds of Twitter members converged to express their amazement that the Titanic was a real ship, with something as hugely significant as the stock market? On the other hand, it’s really not that simple.

If a random Twitter account had started this whole mess it’d be much easier to point fingers and place blame. However, the tweet that sparked this event was sent out through the account of the Associated Press, the people who we rely on for most of our news. These trade-making programs don’t accept just any news story, but if it seems reputable enough they’ll include it. The program makes this judgment call based on a complex algorithm, and the huge majority of the time its findings are right on the money. This time they weren’t, but how could the program have known. The story seemed airtight. It was reported by the Associated Press, a widely trusted organization depended upon by news organizations across the country (and even the world), it was retweeted by several noteworthy and generally trustworthy journalists and it got a huge response, with almost 4,000 retweets within the span of a few minutes. If you’re a trader and you see information like that come across the wire you make a move fast. There’s simply no reason why you wouldn’t.

In the end, the question really is this: should these traders have trusted Twitter in the first place? As a tool to aid successful trading, Twitter seems to good to be true on the surface. It provides short, straightforward bites of information in real time from potentially thousands of sources at once. But the difference between getting your information from Twitter as opposed to getting it from some business publication is that generally before you submit something to a publication of any time an editor has to read it and check the facts contained therein. That’s not necessarily true of tweets. We’d like to think it is, particularly when it comes to the tweets of hugely important organizations like the Associated Press, but the fact is that there’s no way to guarantee that. Of course, factual accuracy isn’t a guarantee in any medium, but it’s much easier to publish a fake tweet than a fake news story. The fact is that Twitter gets hacked far more often than any of us would like to believe. Here are a few of the more notable hacks perpetrated within the last year:

  • In February, Burger King’s hacked account said that they’d been bought by McDonalds, saying (sic) “FREDOM IS FAILURE.”
  • Jeep’s account was hacked as well, causing the company to report that it had been sold to Cadillac and to explain that their name was actually an acronym which stood for “Just Empty Every Pocket.”
  • In September, the Syrian Electronic Army (or SEA) essentially shut down Twitter for about 90 minutes, making the site virtually inaccessible to most users.
  • In October, the SEA altered a tweet made from Barack Obama’s twitter account. The text of the tweet read “Immigration is a bipartisan issue.” This statement was followed by a link which was meant to go to a related article by the Washington Post. Instead, the link was rerouted by the SEA to go to a page about the terrorism-related problems faced by Syria. Following the hack, the SEA sent out a tweet of their own reading “We accessed many Obama campaign email accounts to assess his terrorism capabilities. They are quite high. #SEA”.

Considering these hacks, Twitter is not the secure news source we all want it to be. Following the events of April meetings we called to decide whether or not we should guard against social media’s potential impact on the stock market. What did they decide? Well, let’s put it this way: four months later billionaire investor Carl Icahn sent out a single tweet saying that he though the company was “extremely undervalued.” A few minutes later, Apple’s stock was up by a gain of approximately $17 billon. And if you think Twitter is the only social media site that traders are watching, consider this: in July, Reed Hastings, the CEO of popular video streaming site Netflix posted a status on his personal Facebook page containing the following statement: “Netflix monthly viewing exceeded 1 billion hours for the first time ever in June.” Soon after that status was posted the value of each Netflix share increased by more than $11 to a final price of $81.72. Clearly, the influence of social media has not lessened since April.

The fact remains, no matter how scary it might be to some (me included), that social media does have a huge influence on the stock market, and it will likely continue to for the foreseeable future. Perhaps the oddest thing to consider in all of this is that Twitter can actually be used to predict the stock market, though in a different way than it is currently being utilized. That’s according to the findings of Indiana University’s Johan Bollen, who analyze 9.7 million tweets by 2.7 million users between March and December of 2008 and found that the level of calmness contained therein (as determined by an algorithm called the Google-Profile of Mood States, or GPOMS) could predict the movements of the stock market 2-6 days later with an fairly astounding accuracy of 87.6%.

Of course, there are some issues with the study. 2-6 days is a pretty big window, big enough that skeptics could believe it all to be nothing more than coincidence. Plus, for some reason Bollen and his colleagues included 2.7 million random users in their study, including many that aren’t even based in America, and while some teenager in Australia might have some tangential impact on some unimaginably tiny fraction of the US economy the study would have certainly been better served by limiting results to US citizens only. Regardless, it’s undeniable that, if all goes well, social media sites can be a great tool for investors, but it’s also undeniable that they could very easily be a detriment as well as an aid. Who knows how much worse the situation might be next time something like the April “flash crash” occurs. I guess we’ll just have to wait and see.