The issue of FX liquidity provision has come into sharp focus in recent months. Zeke Vince, Head of eFX Sales and Tan Phull, Head of FX Algos, explain how Bank of America’s (“BofA”)’ approach helps FX algo users to successfully navigate this challenging landscape.
Prior to the outbreak of COVID-19, most clients had become accustomed to a very low volatility trading environment where spreads were extremely tight. Yet according to Vince, as volatility picked up transaction costs started to rise. In addition, “the proposition of saving spread and reducing market impact by using an algo became much more appealing,” he explains. “For that reason, there was a significant increase in algo usage, including those who had previously been less frequent users.”
Now, almost eight months after a clear inflection point in the market, clients and providers are still feeling the impact. Firm spreads in liquid pairs in larger sizes are in some cases still 50%+ wider than the first two months of this year, Phull adds. “In this challenging environment we’ve found an opportunity to step up for our clients and consistently provide access to unique liquidity. Clients are pleased with how we have helped them navigate this difficult period, with the support of our algo and principal market making businesses,” he says.
How algos capture liquidity is also an essential consideration for clients. Liquidity access is at the heart of FX algos and that capture often depends on the franchise of the broader firm, Vince explains. He adds, “At BofA our breadth of larger sized real money, corporates and hedge funds, as well as our industry-leading FX payments business, provide clients with an opportunity to cross both directional and non-directional flows with minimal external market.”
However, there can be certain issues involved with respect to whether an algo uses internal versus external liquidity. With any type of liquidity, the question should be about market impact, warns Vince. “Crossing half or a full spread to interact with internal liquidity may in most cases have lower market impact, but full transparency should be the key,” Vince says.
Next-gen algo performance
Phull adds that clients sometimes will also wish to modify the liquidity pools being used. He explains that allowing customers to have flexibility on routing is a positive development as the venue landscape is so dynamic. “With more and more clients adopting algos, both new and existing secondary venues found a stronger case for catering and pitching directly to the eventual end users of their platforms,” Phull says. “We see this as net positive because clients are learning the nuances and benefits of the different venues that they can trade on.”
Looking ahead, BofA is now working on a customization framework which will give users the ability to tailor highly bespoke behaviors to suit their execution profiles. The user will be able to pick the base strategy and the strategy will pick the ‘flavours’ of execution. “This customization framework will provide the basis for the new generation of re-enforcement learning algorithms,” Vince says. Such changes might be to speed up or slow down the execution depending on market conditions, or even change the skew of participation to be front or more back loaded, he explains. BofA’s aim is also to be as transparent as possible about the liquidity that the algorithms are accessing. “We share information with our clients as the landscape changes and we tune our algos accordingly. This translates to consistent, low impact liquidity and best–in-class algo performance,” Phull concludes.