By John McGrath, Chief Revenue Officer, BidFX
MiFID II regulations were a seismic shift for financial markets when they made buy-side institutions liable for their own execution quality. That change drove the need for a deeper liquidity pool and the best algorithm to execute within it. Cue the rise of the FX “Algo Wheel,” which delivers the winning combination of higher execution quality at lower cost and eliminates bias as to counterparty.
FX algorithms account for up to 20% of global FX spot trading, according to an October 2020 study by the Bank for International Settlements (BIS). Equity markets already surpassed 50%, so that number will only rise, with a recent survey by TradeTech showing 60% of respondents expect usage to increase.
But as demand rises, so does supply, and not all wheels are created equal, especially as some institutions may decide to build in-house rather than buy an off-the-shelf product.
Reinventing the Wheel?
The TradeTech survey showed almost one third of respondents literally intend to reinvent the wheel by building their own product. This is usually driven by a fear of revealing proprietary data, but also in the belief it will be cheaper and better.
However, building is more like purchasing a Scottish Castle: cheap to buy but costs a fortune to maintain. Apart from initial set-up costs, challenges include compatibility with external systems, as well as internal OMS and EMS; back-testing and back-up; constant software upkeep and upgrades; and expanding the product suite to keep up with market developments.
Then there is the risk that the more a firm customises its system for its own use, the harder it is to benchmark for want of true comparison. It may also suffer from its individual trading flow generating insufficient data for statistically reliable conclusions, which is a problem when the buy-side must now undertake “regular and effective monitoring” of their trading tools.
The survey again supports buy over build when revealing clients think bigger is better. When asked about trading priorities, response time and spread were deemed to be twice as important as price, which favours a deeper liquidity pool only the best platforms can provide. The result may be eventual capitulation—join the 54% who already favour buying as the better option—and then try to catch up.
NDF Volume Catching Up Fast
The growing acceptance of Algo Wheels has improved efficiency and lowered operational risk and execution costs, resulting in an increase in “low-touch” trading more reliant on automation. By contrast, “high-touch” flow in less liquid currencies such as NDFs still involves buy-side orders going to sell-side dealers with a limit price range to work on, so have become a disproportionate percentage of total trading costs.
While more than half the respondents suggest NDFs will need at least three years to catch up with the majors, platform providers are more confident. And they have the data.
One of the leading FX platforms reported almost half of their providers added NDFs to their product suite in the first half of the year, prompting a 500% surge in NDF algo volume. That progress will make illiquid currencies much easier to manage, which will only encourage even greater volume.
In addition, newer algos are constantly getting smarter and can respond faster to real-time changes in market conditions. Their data-driven feedback loops and machine learning can also make recommendations on how to execute a particular trade, providing a layer of analysis that will only make an Algo Wheel a necessity rather than a luxury.
From Regulatory Solution to Indispensable Tool
Prodded by the requirement for the buy-side to disclose provable best execution, institutions looked to technology and the benefits of an FX Algo Wheel. Already, that regulatory solution has become an indispensable tool that enhances operational efficiency, reduces costs, improves performance, and learns while doing it. The next survey by TradeTech may well show that the 40% who didn’t expect an increase in the use of Algo Wheels have gone and bought one.
