Algorithms Helped FX Markets Function During Pandemic, BIS Says
(Bloomberg) -- Algorithmic trading for currencies seemed to have helped markets navigate the intense volatility during the pandemic’s outbreak in March.
The Bank for International Settlements made that conclusion in a study on the computer programs that automate about $300 billion of currency trades daily. While algo trading may trigger sharp price moves in crisis times through “self-reinforcing feedback loops,” initial observations from the pandemic showed the systems instead aided getting trades done, the report found.
The BIS, the Swiss-based club of the world’s largest central banks, devoted particular focus to March, when measures of currency volatility more than quadrupled from a year earlier and investors fled to the dollar. Algos contributed to “relatively good two-way order flow” as traditional market makers were less inclined to take more risk onto their balance sheets.
“While many observers and market participants expected them to be less widely used during a crisis, it appears that the opposite has happened,” the report found.
Still, one can’t necessarily draw broader conclusions from this single episode of heightened volatility, the report said. “It is not yet clear what contributed to the more favorable market dynamics in this episode of high volatility, in comparison with, for example, recently observed flash events, which caused some electronic pricing and execution to stop,” the authors said.
Algorithmic trading also contributes to changing liquidity dynamics, transfers risk from dealers to end users, and raises the bar for participants to navigate markets successfully. Overall, however, it helps market functioning, the authors concluded.
The paper shows the findings of a group established in mid-2019 and led by Andrea Maechler, member of the governing board of the Swiss National Bank, to study the implications of the rising use of execution algorithms. It includes survey responses from 70 “sophisticated” global market participants.
©2020 Bloomberg L.P.