Reigning in the Robo-Trader


It's like a Greek myth or a scene from Disney's The Sorcerer's Apprentice: What happens when human beings create machines that can outrun their flesh and blood creators? Someone better be minding the robots.

It's happening today. In the 1990s, trading went electronic and the machines began exchanging the the actual stocks, bonds and options instead of hyper-active men in funny colored jackets gesticulating in a trading pit. Even though it felt like the future had arrived, even more automation was on its way. A few short years ago, investment firms introduced algorithms to trade automatically at unbelievable speeds. When the market moved one way, these complex trading formulas pounced to take advantage of the new market climate.  

A bright and shiny vision of the future?

Not so fast. The same technology that allowed banks to make fast and profitable trades also brought their own unique problems and disasters. After spectacular flameouts like the Flash Crash of 2010 and Knight Capital's rogue algo that lost the respected firm more than $440 million in a single trading day this past summer, Congress wants to bring some sense of order to this brave new world of sub-second automated trading. While this makes sense the regulators claim that they do not have the funds or the technology to do their job and Wall Street is not rushing to foot the bill even if the new laws might help them from imploding much like the credit crisis of 2008. In short, the regulators don't have the data to oversee the markets in real time and Wall Street is checking its fingernails.

It's ironic that one of the most high-powered and profitable industries on the globe has grown so fast, complex and far-reaching that keeping up to manage the industry is nearly impossible with an anemic budget approved by Congress.

What does this have to do with CIOs, you ask? Quite a bit, actually. When the next Flash Crash, rogue algorithm or IPO stumble (Nasdaq with Facebook's IPO and BATS with their own) goes down, every IT chief will have to assure the corner office and the shareholders that they have a plan in place. If you're not stress testing your algorithms for every conceivable what-if scenario, you're playing with fire. Likewise, if the SEC or the CFTC wants your data to explain a suspicious trade, you need that data at your fingertips.

Yes, the machines have never been faster and saying that you and your IT team are trying to catch up is no longer an acceptable excuse. You've been warned.