.svg)
The rapid growth of electronic payments has increased the need for better sanctions screening. As interest in new technologies grows, the market is shifting its approach from “detecting” a suspicious payment to “preventing” it, experts said at Sibos 2021.
Speaking at the Sibos Smarter Screening session, Robert Dean, head of financial crime technology at KPMG, said that the industry is beginning to show a newly forming interest around network-type analytics and contextual pieces which, supported by ISO 20022 data-rich messages, could move sanction screening from detection to prevention.
Sanctions screening is currently going through two transformative events, Matthew Hobbs, head of financial crime strategy and sanctions screening programme at SWIFT, said.
The payments industry is going through tremendous growth. Digital products that previously took years of development before going to market are now implemented “almost overnight.” This brings with it the challenge of understanding and assessing financial crime risks associated with those products and being able to mitigate those risks.
“If we do not keep pace with what is going on in the payments market, it could lead to products being implemented that are vehicles upon which criminals can prosper,” Hobbs warned.
In addition, from November 2022, SWIFT will start adopting the international messaging standard ISO 20022, which will inevitably cause an impact on sanctions screening, in terms of both opportunities and challenges.
In recent years, industry players have started to treat this common, global standard not just as an implementation obligation but also as an opportunity.
ISO 20022 has the potential to have a key impact both on sanctions screening and transparency.
Due to the structure of the ISO 20022 message, it allows organisations to make targeted screening and look for a specific piece of information in specific fields that “should reduce operational overheads of screening as we see it today,” Dean explained.
Transparency can be another benefit of the messaging format. Under FATF Recommendation 16, “countries should ensure that financial institutions include required and accurate originator information, and required beneficiary information, on wire transfers and related messages, and that the information remains with the wire transfer or related message throughout the payment chain.”
Although the recommendation has been out for years, many organisations are still struggling to fully comply with the “spirit” of the recommendation, according to Dean.
The data fields available in the new messaging format will eventually help organisations to make their messages and payments as transparent as they possibly can. With that said, Dean gave the audience a note of caution regarding reliance on the data transferred.
It is brilliant to have a message format, structures and data dictionaries, “but getting to the point when you can rely upon it is probably one of those things that I can still see being a little bit of challenge as the industry adopts,” he said.
Supporting Machine Learning
In addition to shifting to the new messaging format, financial institutions have been showing an increasing interest in the use of machine learning in sanctions screening, with many industry players having heavy investment in the technology, Dean noted.
Machine learning has the potential to narrow the risk in the system by reducing the noise and getting to the risk quicker, Richard Norman, head of financial crime screening operations at Barclays, added.
Machine learning in payments is now taking a step further by allowing risk scoring not just to rely on data but also on interpretations between different data points, such as historic decision making.
Based on a newly forming interest in reverse engineering of some of the features, data sets, and techniques within transactions screening, Dean believes that organisations may build these elements into their payments origination system instead of building them into the detection capability of a filter.
An interest is beginning to form around network-type analytics and contextual pieces pulling further astray from the payment process, he added.
“It is always better to prevent the origination of a payment, which potentially has a sanctions nexus to it, than hope that your filter will capture it before it exits the organisation.”
“As you move through the ISO 20022 journey, and you begin to change the systems and infrastructures you have to support that, the opportunity exists to really begin to look at what could we do to make a step to move from protection into prevention,” Dean said.
