Max Bowie: Hacked Off: News Scandals and the Future of Trading

If you’re a trader who incorporates news signals into your strategy or uses news as a reference point to support trading decisions in any way, or if you’re someone in market data management who serves such traders, then some recent developments may make you think twice about placing too much faith in the news flowing into your trading models, and hence in the output of those models.

What’s my interest in headlines, apart from our own? News is as legitimate an input to trading as price data. If you know—or can predict with reasonable accuracy—what news is coming out, and when, you can anticipate how the price of an asset will move. And many firms have tried to incorporate this, enlisting technology vendors who help develop newsflow algorithms to predict the impact of news based on historical analysis, and who quantify the sentiment of that news to predict exactly how much a price will move.

New Channels
And since news is increasingly breaking first through social media channels such as Twitter, firms and vendors are finding ways to harness this as a new data stream. However, although traders may be discerning about whose tweets they base trading decisions on, hackers recently fired a warning shot across their bows by taking control of the Associated Press’ Twitter feed and tweeting that the White House had been bombed and US President Barack Obama was injured. The news wasn’t true, and was quickly corrected by the AP, though not before stock prices dropped then rebounded in what became known as the “Hash Crash.”

The AP has also been embroiled in another scandal beyond its control: the secret seizure of its reporters’ phone records by the US Department of Justice, supposedly to root out leaks that cut short a CIA anti-terror operation. Aside from being the kind of intrusion into the free press usually only seen under much scarier regimes and in violation of the First Amendment, this is important for those who trade the news because any important news that gives an early warning of something that could impact price movements is generally not officially sanctioned, and depends on confidential sources, who are often motivated by a belief that they are doing the right thing by making information public. If these people no longer feel confident that they can provide this information without being identified, they’ll stop doing it. And that won’t just leave us journalists at a disadvantage: Fewer sources means lower-quality news for traders to base decisions on—or, for their algorithms to base decisions on. And if there’s one thing that an algorithm needs to function properly, it’s high-quality data inputs.

Scandal
But another news scandal has traders thinking twice about their data. The revelation that Bloomberg reporters were able to access some information on the usage habits of its terminal clients has left firms uncomfortable about the vendor’s level of insight into their activities. Though Bloomberg says it has curtailed the process, and that its reporters only had limited access to user information, firms remain unconvinced, and are reportedly investigating other vendor solutions to provide services such as messaging, to reassure themselves that Bloomberg isn’t eavesdropping on any other information that passes across its systems. With firms so dependent on privacy for the success of their trading strategies, and paranoid about giving away details of their strategies to their competitors—who could use the information to trade against them—the last thing they want to do is give anything away through a third party.

If sources no longer feel they can provide information without being identified, they’ll stop doing it, which means lower-quality news for traders—and their algos—to base decisions on.

If your strategies are designed to exploit the impact of “herd” movement as a result of news, then you’re feeling good about these developments. But if you don’t have a crystal ball and depend on news being reliable and accurate, you’re probably feeling nervous about your news. Perhaps the next development will be a tool that independently verifies a story’s accuracy and provides a measure of confidence in each news item, so traders can ascribe a specific confidence requirement for each input before an algorithm can act on them. Or perhaps you’ve already invented it? If so, we’d love to hear from you.