Juliette Kennel and Sam Romilly
Juliette Kennel and Sam Romilly

SWIFT works to unlock the value from its FX data

Robust and reliable data is a core requirement for participation in FX markets at any level and in any capacity. But where can such data be sourced in sufficient depth and in real-enough time to hold its value through any transactional or analytic requirement? e-Forex spoke to Juliette Kennel, Head of Securities and FX Markets, SWIFT, and Sam Romilly, Head of FX New Business Development, Capital Market, SWIFT, who can answer that question. 

SWIFT is uniquely placed to provide fact-based data products and market insights. The SWIFT network is used directly to confirm FX trades by 8,000 financial institutions who are connected to the network directly, and indirectly by an additional 2,000 corporates and investment managers.

These FX trade confirmations use the SWIFT message types MT300 and MT305, of which more than 1 million are exchanged each day. Juliette Kennel says: “We estimate that only around 10% of the total value confirmed by these messages is also settled via CLS, meaning that the rest of our dataset is unique to SWIFT and not available from any other source.”
SWIFT also has a significant FX dataset related to transactional banking that allows an understanding of corporate sector flows. That comprises FX trades arising where clients of the transaction banks issue payment instructions that require a currency conversion. Kennel says: “These FX trades are largely managed though the use of a bank’s in-house accounts so they don’t go to the market.  However, the resulting payment instructions are sent over SWIFT in the form of MT103 messages.”

There is, of course, strong governance in place regarding the protection and confidentiality of SWIFT data. Kennel says: “The data belongs to our member-bank community and SWIFT operates under strong principles that govern every aspect of our data services. Over the years and responding to user requests, the Board has approved the extraction of message data fields that have enabled SWIFT to develop a range of Business Intelligence (BI) tools to allow SWIFT members to better understand their business. This data is made available to our members, who have access to their own data, the data on the total market, and peer benchmarking on an anonymised and carefully protected basis.”

Integration and analysis

Data of such scope and quality is an asset to be used. Sam Romilly says: “Like all companies, SWIFT sees the potential for improving services and meeting customers’ needs through the greater use of data. Our members are now looking for new datasets to help remain competitive and to manage risk. The SWIFT dataset is not limited to just the FX industry and our members have realised it can complement their own datasets to help them realise their own strategic and commercial objectives.”

Latest technology has been used to extract the data and process the data in a secure and confidential manner. Romilly says: “We also use a state-of-the-art tool to visualise the data for the customer who subscribes to our Watch Analytic products. However, the more advanced new technologies around AI and quantitative analysis are what our member clients would use to integrate and analyse our datasets together with their own data pools.”

In 2017, SWIFT launched an FX peer-benchmarking service, which provides a view of an institution’s ranking on a country, regional or global basis compared to anonymised peers across multiple permutations of segments, currencies and flows. This report was enhanced in 2018 with more granular data that identified the actual FX instrument types. Romilly adds: “This year we have worked with selected clients to identify the full value inherent in the FX dataset. We have been able to align our data with the clients’ own business flows and come up with a common agreed set of assumptions regarding segments and definitions around how to classify and count FX instruments.”

There’s been significant interest. SWIFT’s new data range promises to: help SWIFT users manage their business risk; help SWIFT users understand their competitive position; help the industry to better understand economic trends. Not surprisingly, SWIFT users have been quick to respond. Romilly says: “Data provided by SWIFT is perceived today as a valuable tool to analyse and forecast important economic trends. Studies conducted by SWIFT and shared with the community allow us to point to correlations between SWIFT traffic and markets’ volatility, payment flows and interbank lending.”

SWIFT’s data is seen as special because it is based on actual transactional data and not surveys
SWIFT’s data is seen as special because it is based on actual transactional data and not surveys

SWIFT’s data can also prompt analysis. Romilly continues: “As with any dataset no-one really knows what they want to see in advance.  It is only by looking at trends and movements that exceptions can be spotted and then deep dives into the data granularity can provide answers.  This could be answers to questions such as ‘Why did we lose market share in this period, or this currency, or this client segment?” Comparisons of an institution’s own activities to what is seen happening in the total market can help drive commercial and strategic decisions.”

Towards a data-rich future

SWIFT’s data is seen as special because it is based on actual transactional data and not surveys, and because it comes from an independent source. Kennel says: “Overall the feedback is positive and customers see value in our dataset.”

Appropriately, the development of the new service was collaborative with members from the outset. Kennel continues: “SWIFT set up a pilot project with five banks; we called them our Design Partners.    They were provided with samples of the SWIFT FX dataset in return for their advice and ideas around how to enhance, process and present the data.”

For the first time, it is now possible for SWIFT to show the relative percentage breakdown of instrument type for the global FX market. Kennel says: “The significance of this is that this data is now extremely useful and equivalent to the data held by each of our members. Before, when the individual instrument types could not be classified, it was at best just interesting.”

Will SWIFT’s data services be developed further in coming months? Yes. Kennel says: “There are three directions we will be following. The first is to make sure our data delivery is scaleable and industrialised so that we can meet new client demand. The second is to continue to evolve our understanding of the value of the dataset and seek new ways to slice the data to reveal more of its value. And the third is to explore in the long term ways we could potentially work with data aggregators within the limits of our governance policy to provide new combinations of datasets and data services to our members.”