Are Artificial Intelligence (AI) programs more efficient than traditional buying and selling methods? JPMorgan is the latest bank to pursue research into this field, developing a first generation bot to execute trades across its global equities algorithm business. Per the group’s latest trial however, the AI utility known as LOXM was shown to demonstrate improved efficiency over existing means, per an FT report.
Exactly what this means for the broader equities industry is simply a matter of conjecture at this juncture. Robotic trading or the use of AI rather has long been seen as a natural evolution amongst financial firms. However, humans have retained an integral focus at all major venues even as the industry moves towards more algo-focused trading regimes.
LOXM is in control
JPMorgan had been developing its own AI, LOXM, which was given a trial run for its European equities segment. More specifically, the AI has been live since Q1 2017 with early performance analysis yielding positive returns. The performance has evidently been so successful that the lender is green lighting its launch for Q4 in Asia and the US across its platform.
Indeed, David Fellah of JPMorgan’s European Equity Quant Research team, commented: “Such customisation was previously implemented by humans, but now the AI machine is able to do it on a much larger and more efficient scale.”
Many other lenders have been tinkering with AI capabilities as well, though a Q4 launch would make JPMorgan one of the industry’s first to go live. The move is an endorsement for LOXM, whose central role is to execute client orders with optimized speed at the best price.
Cost efficiency driving innovation
This has been achieved through machine learning techniques, i.e. using lessons it has learnt from billions of past trades on a simulated and real-time basis. Consequently, the AI has learned how to effectively reconcile a range of issues, including avoiding market price movements via the jettisoning of large equity stakes.
Ironically, the investments in LOXM as well as other AI initiatives were originally designed to help curtail costs for existing functions at investment banks. While hardly the only pioneer in this field, JPMorgan’s efforts and research appears to be paying early dividends, which can help cut down tremendously on multiple time consuming human-designated tasks, such as client post-trade allocation requests, among others.
A broader industry adoption could also yield a windfall for other banking segments, including gains across automatic hedging, market making, and others.
Be First to Comment