Key Takeaways
- • JP Morgan's futures algo volumes grew 40% year-on-year since 2016, and algos now make up almost 20% of the bank's total futures trading flow, up from roughly 4-5% in 2016 and 2017, according to figures seen by The TRADE.
- • Customised algorithms have roughly tripled in volume each year over the past three years, with close to 50 customisations in 2020, up 21% year-on-year and from close to zero in 2017.
- • JP Morgan's Aqua algorithm now uses reinforcement learning, a form of machine learning, to generate order routing and placement signals, and its Roll Algo model went live for the February 2020 US Treasury roll.
- • Peter Ward, global head of futures and options electronic execution at JP Morgan, expects buy-side adoption of futures algos, particularly customised algos, to keep accelerating in 2021.
- • Rival Citi launched a suite of futures execution algorithms in January 2020, with its electronic traders handling customisations on clients' behalf.
Summary
Adoption of futures algorithmic trading at JP Morgan accelerated sharply in 2020 as buy-side traders globally turned to the bank's machine learning-enabled algorithms amid intense market volatility. Speaking to The TRADE, Peter Ward, global head of futures and options electronic execution at JP Morgan, said volatility contributed to recent growth but that adoption had picked up pace with clients over the last several years. According to figures seen by The TRADE, futures volumes traded via algos at the bank have increased 40% year-on-year since 2016, and algos now account for almost 20% of total futures trading flow, compared with roughly 4-5% in 2016 and 2017.
Customised algorithms have been a particular driver. Volumes on customised algos have roughly tripled in each of the past three years, and the number of custom algos rose 21% in 2020 to almost 50, from close to zero in 2017. The bank's flagship liquidity-seeking algorithm, Aqua, is the most common foundation for client customisations. JP Morgan rebuilt its algo platform around five years ago to give the buy-side more control over parameters, and its electronic traders can configure further customisations on clients' behalf. Ward said that in the past, customisation required developers to code and release changes, which was time-consuming, and that the bank now offers a scalable approach.
Aqua has also been the focus of reinforcement learning work. The technology allows the algorithm to learn from its previously generated signals and evaluate performance over time, with the signals determining whether the algorithm crosses the market or remains passive. The first application was a model focused on quarterly futures roll dates, which JP Morgan branded the Roll Algo. It went live for the most recent US Treasury roll in February 2020 and was especially popular with buy-side traders, according to Ward, who said the algorithm learned from that period and performance in the next quarterly roll should improve. The bank has also developed Target, which trades around the cash or futures close, Multi Leg Strategy for trading multiple instruments across futures and US Treasuries, and options algos. Options on futures volumes surged in 2020 after floor closures at exchanges such as CME shifted liquidity to electronic channels.
Ward attributed the trend to explicit best-execution requirements and buy-side interest in addressing futures and options market structure challenges, along with extended futures trading hours and fragmented liquidity across venues. He noted that the buy-side client base has broadened from a small number of large hedge funds to pension funds, asset managers and insurers. Citi rolled out its own futures execution algorithms, including its Arrival strategy, across major exchanges in January 2020, but its electronic traders handle all customisations on clients' behalf, according to Gordon Ball, head of EMEA futures electronic execution at Citi. Ward expects the pace of adoption to continue in 2021.
Customised algorithms have been a particular driver. Volumes on customised algos have roughly tripled in each of the past three years, and the number of custom algos rose 21% in 2020 to almost 50, from close to zero in 2017. The bank's flagship liquidity-seeking algorithm, Aqua, is the most common foundation for client customisations. JP Morgan rebuilt its algo platform around five years ago to give the buy-side more control over parameters, and its electronic traders can configure further customisations on clients' behalf. Ward said that in the past, customisation required developers to code and release changes, which was time-consuming, and that the bank now offers a scalable approach.
Aqua has also been the focus of reinforcement learning work. The technology allows the algorithm to learn from its previously generated signals and evaluate performance over time, with the signals determining whether the algorithm crosses the market or remains passive. The first application was a model focused on quarterly futures roll dates, which JP Morgan branded the Roll Algo. It went live for the most recent US Treasury roll in February 2020 and was especially popular with buy-side traders, according to Ward, who said the algorithm learned from that period and performance in the next quarterly roll should improve. The bank has also developed Target, which trades around the cash or futures close, Multi Leg Strategy for trading multiple instruments across futures and US Treasuries, and options algos. Options on futures volumes surged in 2020 after floor closures at exchanges such as CME shifted liquidity to electronic channels.
Ward attributed the trend to explicit best-execution requirements and buy-side interest in addressing futures and options market structure challenges, along with extended futures trading hours and fragmented liquidity across venues. He noted that the buy-side client base has broadened from a small number of large hedge funds to pension funds, asset managers and insurers. Citi rolled out its own futures execution algorithms, including its Arrival strategy, across major exchanges in January 2020, but its electronic traders handle all customisations on clients' behalf, according to Gordon Ball, head of EMEA futures electronic execution at Citi. Ward expects the pace of adoption to continue in 2021.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.