JP Morgan Asset Management’s equity trading team has developed a new model using machine learning to make the execution of orders more efficient and less expensive.
The proprietary model was created by quantitative analysts and traders on the equity desk, led by Lee Bray, head of equity trading for Asia Pacific at JP Morgan Asset Management, using data patterns to find the best execution strategy for trading orders.
The artificial intelligence model constantly learns the best outcomes for trading orders and adapts as market conditions change and new data is available, according to the firm.
Using these patterns and algorithms, the model can target the probability of the best performance, then auto-route and execute orders accordingly.
“To develop the model with machine learning, we tapped into techniques more commonly found at companies like Facebook and Google,” Bray said.
“By creating a systematic, adaptive model able to alter actions based on mathematical patterns rather than relying on human input, we’re transitioning equity trading to be more scientific and quantifiable.”
JP Morgan added that currently the model provides recommendations to human traders, but it is increasingly taking over an automated role in executing transactions.
The Asia Pacific equity trading team at JP Morgan Asset Management is now targeting to have around 50% of the notional value of all regional trading activity automated with the machine learning model by the of this year.
“With myriad options available for executing any given order, particularly smaller or more routine orders, an intelligent model can identify the best execution more efficiently than a human,” Bray concluded.
JP Morgan Asset Management said it invested significant resources in building machine learning tools to improve the global equity trading business.