Skip to main content
Header logo

The global provider
of secure financial messaging services

Technology is cutting costs but cleaner data could transform how FX markets work

Technology is cutting costs but cleaner data could transform how FX markets work

10 December 2019

Digital technology now promises full automation for FX transaction processing, but the power of AI and machine learning is yet to be felt because FX datasets are not available in standardised, machine-readable formats.

This is one of a series of articles based on panel discussions at Sibos 2019, where industry figures came together to examine the issues impacting FX markets.

Technology has already had a major impact on buy-side activity in FX. Electronic trading platforms continue to proliferate, while algorithms, liquidity aggregation tools and transaction cost analysis (TCA) services are changing how corporates and asset managers discover prices and assume, manage and transfer currency risks.

The complete automation of FX trades, from execution, through settlement, to reporting, without any human intervention at all, is now possible.

David Leigh, head of electronic FX trading at Deutsche Bank, said the bank was already offering corporates executable prices in real time across 125 currencies, including restricted ones.

“One of the things we’ve been working on is taking a wholesale institutional streaming FX price and plugging that directly into a client’s ERP system,” he said, “so that they are actually able to use that for cross-border payments, which maybe today are being processed at end-of-day against fixing rates.”

FX Day Sibos 2019 London

Streamline processes through automation

An important operational benefit is the elimination of post-trade reconciliation, by corporates, of expected and actual FX exchange rates, because they are the same.

Christian Mnich, head of solution management at SAP, agreed that “seamless connectivity and APIs” are enabling corporate treasurers to centralise FX transactions and automate spot and forward execution at banks and trading platforms.

A FinTech provider, Kantox, now provides a service that enables corporates to automate execution by linking their ERP systems to banks and trading venues via Application Programme Interfaces (APIs). “We replace manual execution with a machine that works for 24 hours a day, and makes execution completely streamlined,” explained Philippe Gelis, CEO of Kantox.

Fintechs as drivers for change

The application of new technologies is a huge driver of change for the FX industry.  Fintechs are targeting the FX market particularly with alternatives to current execution and post-trade processing. Four fintechs had the opportunity to pitch their solutions to the Sibos audience for their review.

One of these was Cobalt who look to truncate the processing of post-trade functions. By capturing all the details of an FX trade once, Cobalt is able to collapse matching, confirmation, allocation, aggregation, netting, compression, novation, switching, credit allocation, payment instruction and reporting into a single, seamless process.

Adrian Patten, co-founder and chairman at Cobalt, said one sell-side client expects to save US$25 million a year within three years simply by retiring legacy post-trade services and technologies.

Another, Siege FX, promised end-investors lower execution costs through innovation at the pre rather than the trade or post-trade stage. It aims to achieve this via anonymous and invisible peer-to-peer netting of FX transactions between end-investors.

The two other FinTechs proposed reduced risk through increased speed. Baton for example uses cloud-based software to settle currency transactions through pre-funded accounts in commercial bank money within three minutes. Koine offers a similar service using digital tokens linked to pre-funded accounts.

FX Day Sibos 2019 London

Artificial Intelligence (AI) and Machine Learning (ML)

The FX industry is already looking to benefit from  the exciting new technologies associated with Artificial Intelligence (AI) and Machine Learning (ML) but the benefits depend on having the right data.

Chris Purves of the FRC Strategic Development Lab at UBS pointed out that a lack of standardised, machine-readable data was inhibiting the impact of AI and ML on bank trading desks and trading platforms (though both are experimenting with it).

He predicted that AI and ML would, in time, help traders identify the best sources of liquidity, direct sales calls more precisely and replace the - presently manual - process of allocating credit.

Seed Amen, a quantitative currency trader and founder of trading strategists Cuemacro completely agreed "The key point is to try and get data in machine-readable formats, so, if banks, to give a simple example, publish PDFs, maybe they could also give that to their clients in a machine-readable format as well. I don’t know whether that is something that can be mandated specifically, but it is something that maybe clients of institutions can encourage them to do ,to say, ‘Look, we want to have all the data that you have in a machine-readable format because then it is easier to ingest, easier to process and easier to read.”

Can regulators keep up with new technology

Concerns that regulation would hinder the beneficial impact of technology were dismissed by Mark Yallop, chairman of the Fixed Income, Currencies and Commodities (FICC) Market Standards Board (FSMB), who argued that FX market participants could allay the concerns of regulators themselves.

“The issue that people don’t always understand is that there aren’t really any standards setters in the world today who have an agreed point of view on what the solutions are to the problems that new tech is throwing up,” he said. “A world in which national regulators dominate thinking in their own jurisdictions is not well set up to deal with global technology developments. That is why I think private sector standards setters – not just FMSB but others as well – have to step into the gap to address some of these questions.”