The importance of effective algo execution on both the buy- and sell-side was at the forefront of discussions for panellists speaking at TradeTech FX in Barcelona.
Ralf Donner
Discussing how in-house and third-party algos can boost trading strategies, experts at the algos panel were quick to highlight the vital role that these play for FX traders in enhancing efficiency and automating their tasks to achieve faster and more impactful trades.
“There was a comment made that mentioned surprise that central banks are using algos. I view it the opposite way; I’m surprised more people aren’t using algos,” said David Kalita, chief executive of Quantitative Brokers.
“When we look at how an algo operates, it’s a crew of traders in a room looking at a TCA or something going on in the market and thinking about what we would do as a trader in this situation. It allows people to scale and do things much faster and across a lot more markets.”
This sentiment was echoed by Ralf Donner, executive director at Goldman Sachs, who underlined the influence of algos on market structure, and the ability to unlock capabilities not possible for a human trader to execute.
“A human trader can’t have 10 keyboards in front of him to trade in all the different secondaries, it’s not possible. Only a machine can do that,” he asserted.
Similarly, when questioned on how firms evaluate their algo execution, discussions turned towards the importance of liquidity management mechanisms to achieve the greatest algo outcomes.
“Even the most powerful algos cannot reach their expected performance without proper liquidity,” said Sana Horrich, chief FX dealer, Banque de France.
“Liquidity is a backbone of algos and given the very highly fragmented market structure, they allow us to gain access to wider liquidity and connect to the available liquidity in the market.”
She added that continuous interaction with providers is essential to stay on top of the evolution of tools and navigate distinctions between providers, which are often put “in competition” with each other to deliver the most efficient algo offerings.
Algos and AI – a match made in heaven?
As increasingly seen in industry dialogue in recent times, conversations also turned to AI, specifically the inextricably linked role that it plays in FX algo development.
For the panellists, integrating AI into algo construction provides both advantages and challenges, with a key benefit being its capacity to help find the mid-point, a particular pain point for FX traders.
As markets begin to become more fragmented, mid-point finding is an increasingly difficult task across the FX industry, so integrating AI can be useful to find the best short-term predictive price power.
However, those on the panel were also quick to emphasise that there must be a balance when introducing AI to algos, particularly when the technology takes precedent over human intelligence.
Highlighting this, Donner commented: “I can’t be asked how it is possible that someone else was 15 basis points from me on a trade, and I just throw up my hands and say I have no idea, it was AI. That for me is a non-starter.
“I think over time maybe AI can use for certain pieces of algo constructions, as long as it doesn’t detract from the overall development.”