ISO migration for cross-border payments - a wake-up call

Worldwide, payments systems are being migrated to the ISO 20022 standard UNIFI (Universal Financial Industry Message Scheme). In the euro area, this process began with the introduction of SEPA in 2007. Next November, the time will come for high-value and cross-border payments as well. The financial world expects a lot from the migration. In a survey by msgGillardon and Unifits, 99 % of the experts questioned had a highly or rather positive view of the ISO migration. And indeed, the new standard offers many advantages once it is supported as comprehensively as possible.

The topic is anything but new, and yet the road to it is difficult and the introduction had to be postponed several times. In this context, the Eurosystem of the European Central Bank relies on a big-bang approach, while SWIFT envisages a three-year coexistence phase of MT and MX formats for global payments via the SWIFT network. Although the big-bang migration seems like the more difficult project, as all banks in Europe with TARGET2 or EBA EURO1 connections have to switch over at the same time and without any delay tolerance, the coexistence phase of MT and MX also has its pitfalls.

The TARGET MX migration is much more than just a format change. At the same time, technical access will be switched to ESMIG (Eurosystem Single Market Infrastructure Gateway). An ISO-based recall process and central liquidity management (CLM) with new liquidity accounts (MCA/DCA) will be introduced. The ECB has stipulated a test period of around one year for the banks and is continuously monitoring the transition – at quite a few banks, their own TGT MX projects are in yellow or red status. The secure operation of high-value payments in Europe appears to be at risk at some institutions.
However, this wake-up call is for the migration in cross-border payments. The SWIFT initiative CBPR+ (Cross-Border Payments and Reporting Plus) tries to facilitate a soft transition through the 3-year transition period. Each financial institution should be able to meet the goal of the ISO migration at its own pace. As a result, conversion takes place in many places instead of end-to-end processing of the ISO data. Some financial institutions continue to send and receive MT messages for payments, others are already using the ISO formats. Furthermore, the migration takes place in stages per message type anyway.

A closer look shows that the transition period is expected to cause huge problems both for financial institutions that have not yet migrated to ISO and for financial institutions that are technically ISO-ready:
Financial institutions that can only process MT risk slower processes due to conversions and have to invest in, for example, transitional solutions for compliance reasons as in some processes the original ISO formats have to still be retrieved and checked. The worst problem: the threat of losing data such as addresses...

Other problems lie ahead for all SWIFT participants: the MT/MX transition period leads to a multitude of new scenarios and use cases due to the possible formats in messages of a transaction – notification in MT, cover in MX, fee request in MT, recall in MX, etc. – which can be used in different ways. Transaction monitoring becomes more difficult and different mapping at participating financial institutions can lead to misinterpretations.

I would like to briefly outline a particularly critical scenario: a payment ordered in ISO format is executed via a longer chain of correspondent banks, one of which can only send MT formats. The MT correspondent receives the converted ISO payment and is obliged to retrieve the original ISO format, but cannot pass on the truncated data in MT format. The SWIFT Transaction Monitor does not yet contain a "golden copy" of the original payment, and thus the recipient bank can now neither read out or query the lost data from the MT message nor from SWIFT. In this scenario, the information is therefore not completely transported from the ordering party to the recipient, although both the ordering party's bank and the recipient's bank can already process ISO formats.

My recommendation is therefore, despite the acute shortage of expert staff, to make use of the remaining time and test the cross-border payments migration with all its many scenarios (which are actually new in this context). Ideally, you should test together with your correspondents in particular and check whether the mapping of bilaterally agreed MT assignments still fits in the ISO world.
After all, all financial institutions that act as transaction banks should be able to process the ISO data end-to-end without conversion.

Otherwise, international payments will be faced with a three-year long ordeal.

Thomas Riedel