Surge in use of algorithms may hurt quality of FX prices - BIS

LONDON (Reuters) - The growing use of computer-run algorithms to trade currencies is helping foreign exchange markets function but could undermine the quality of forex prices reported publicly if they become too common, the Bank of International Settlements said on Friday.

FILE PHOTO: Saudi riyal, yuan, Turkish lira, pound, U.S. dollar, euro and Jordanian dinar banknotes are seen in this illustration taken January 6, 2020. REUTERS/Dado Ruvic/Illustration/File Photo

Execution algorithms, or “algos” as they are popularly known, are used for an estimated 10%-20% of daily global spot trading.

Regulatory demands to minimise costs and improve transparency have spurred the growth of algos in the $6.6 trillion-a-day foreign exchange market, where trading is fragmented across dozens of different platforms.

The low market volatility seen since the 2008 global financial crisis has also increased take-up, with many participants turning to them as a more efficient way to trade than traditional voice-brokered or electronic methods.

That efficiency was evident during the COVID-19-induced market selloff in March, when currency market volatility exploded after years of calm. Providers reported more than a doubling of algo-related volumes relative to average usage, the BIS said.

“The impact on market functioning will depend on many factors, including the direction in which EAs (execution algorithms) evolve, their market share, how well their risks are understood and managed, and how they interact with other developments in financial markets,” the BIS, an umbrella group for central bankers, said in a report.

FX execution algorithms are automated trading programs designed to buy or sell a predefined amount according to a set of instructions.

They have evolved from simple mechanical rule-based trading to more complex strategies that assess market conditions in real time. Users can bundle order book data across multiple venues and slice large orders into smaller chunks.

Increasing algo usage has also facilitated a growing trend towards internalisation, whereby a market maker -- usually a bank -- will try to offset currency trades internally without passing them to external venues, a process known as “dark” trading.

But the BIS warned that while market participants have been able to benefit from banks matching trades internally, at the extreme too many internalised trades could reduce turnover on primary venues and undermine the quality of the prices they report.

This matters because prices from primary trading venues such as Refinitiv and EBS are used as reference prices for other currency trading platforms and for bilateral trading.

Rising algo usage has also been accompanied by more so-called “flash crashes”, in which prices move sharply for unexplained reasons, that have drawn the attention of global regulators.

(GRAPHIC: FX algo BIS - )

Reporting by Saikat Chatterjee; Editing by Tommy Wilkes and Hugh Lawson