Algo trading software and the human touch: Finding the execution sweet spot

For much of the past decade, market players have been focused on low-touch or no-touch execution solutions. As market electronification has taken off, quant funds and other proprietary trading firms have looked to automate most, and in some cases all, of their trading. The development of ever more sophisticated algo trading software bore out the wisdom of this strategy and you don’t have to look far to find low-touch advocates.

Equities and derivatives trading have led the way, with electronification increasingly the norm. In the case of equities, where trading has become increasingly electronic over the past two decades, it’s now even becoming common for buy-side players to look to build custom algorithms. A survey by Greenwich Associates late last year of sell-side executives showed they expected nearly half of their clients to ask for customised order-handling logic in the coming year.

But even fixed-income trading, which traditionally has been seen as a market that has required human judgment and discretion when it comes to trading, has become more electronic. A separate report by Greenwich, cited by Bloomberg, said 69% of trading in the U.S. Treasury market was through electronic platforms. And the Financial Times recently quoted the new chief operating officer at corporate bond trading platform MarketAxess as predicting an imminent surge of electronic trading in fixed income. “The level of automation will explode in the years ahead,” Chris Concannon, former president of Bats Global Markets, told the FT. In Europe, meanwhile, the best execution requirements in MiFID II are also pushing traders towards electronic trading.

Beyond automated trading

Still, while many firms may have grown comfortable trading on autopilot, there are times when a trader wants to step in and take control of the situation. It may have to do with orders of a certain size. Or it could be based on events when a market has become dislocated or an unusual situation has occurred. Even traders at firms that are heavily algo-reliant are bound to feel the occasional urge to make real-time decisions about how to get in, or out, of the market. It’s not just buy-side traders either. Say a broker has a position still open that hasn’t been picked up by an auto-hedger. Relying on a GUI to tidy things up may be the simplest route.

It’s completely understandable. In fact, it’s worth taking a moment to think about what algorithms are particularly good at versus the kind of tasks that humans can do better. Dr Hannah Fry, in Hello World: How to Be Human in the Age of the Machine, recently wrote of how superb algorithms are at prioritising information, classifying it, filtering it and finding links. But algorithms, she notes, have weak points. Understanding context is one of them.

Fry may be writing for a wide audience, but the same logic applies in the market. An execution algorithm cannot see the reason behind the widening or narrowing of spreads and sometimes that reason makes all the difference. Conceivably, an outbreak of war and an interest rate change could, in the initial few seconds, have the same effect on the market – but longer term, the implications for the market of one versus the other will be very different.

Hybrid algo trading software

In some cases, an execution decision may not be about understanding context. It may simply come down to convenience. Consider a flow trader who has leftover trades to take care of after clients’ orders have been crossed by an internaliser. Or may be a broker wants to interact with the flow to look at the current positions and flatten out exposure. Going high-touch might be the best route.

So, for firms that do want the ability to turn off the autopilot, what sorts of execution system make sense?  A full-option, sophisticated execution management system (EMS) designed for high-touch trading may be overkill. What’s needed is an intuitive front-end that allows a trader to take control quickly and easily on those rare occasions when it’s needed. This is not the sort of system that has unnecessary bells and whistles or would require extensive training in order to use it.

In other words, many firms may find that what they actually need is hybrid algo trading software, one that is designed to allow most of the trading to be automated but will make it easy for those times when traders want to take over the wheel.

As metaphors go, this one is apt. Think of an EMS like a self-driving car. Manufacturers could build one without a steering wheel, but most people would feel a lot more comfortable getting inside knowing one is still there. After all, you never know when you’re going to want to take control – just in case.

__________

Learn more about Vela’s execution solutions. Read this case study on how electronic liquidity providers can rapidly scale their businesses through a multi-broker direct market access platform.