Financial crime: Towards better transaction monitoring

05 April 2018

In our last blog we explored how to use data and technology to understand your customers, for the third blog in our financial crime series we've looked at transaction monitoring systems and how technology and data analytics can help make transaction monitoring more efficient. 

Transaction monitoring – identifying anomalous transactions and suspicious patterns of transactions to and from bank customers – is one of the central requirements of anti-money laundering (AML) legislation. Financial institutions have been required to implement transaction monitoring systems since 1991 when the first AML directive was issued by the EU – but the efficacy of these systems, for a number of reasons, remains open to debate.

When transaction monitoring systems are ineffective, it’s usually for one or more of these common reasons:

  • Missing and/or poor quality data;
  • The use of rules-based scenarios, undermined by inappropriately-set threshold criteria;
  • Inappropriate customer segmentation, or segmentation that isn’t detailed enough to allow for a sensible distinction;
  • Monitoring that doesn’t address the risks posed by particular products or countries where business is carried out; and
  • A disconnect between the teams handling investigations and analytics. Banks aren’t able to learn from their successes or failures – often the case documentation to provide that feedback isn’t detailed enough to provide good evidence to regulators or for internal audit purposes.

When you consider the sheer volume of transactions processed every day – SWIFT, for example, records on average 28.4m FIN messages every day – it’s easy to understand the scale of the challenge.

Typically, monitoring systems are based around simple rules – transactions of more than a certain amount or involving high-risk countries are automatically flagged. But this sort of approach produces an extraordinarily high number of unproductive alerts; many institutions witness false positive rates of over 90%. That can mean investigating, in the case of the largest banks, up to half a million alerts a month.

And that’s where the problems really begin. Investigating transaction alerts is painstaking and often tedious. Traditionally, systems were set up only to trigger and not to gather information that would be useful for an investigation (let alone collect richer intelligence from other sources), so the lack of data attached to alerts means that systems have to be manually trawled for the data that’s required.

The good news is that technology and data analytics have the power to make transaction monitoring far more efficient and effective. This is why many banks are looking closely at Robotic Process Automation (RPA) to help improve transaction alerts for example. But as with any new technology, the tools alone aren’t enough; if banks are to make the most of this next generation of technology, they need to make absolutely sure that their underlying rules-based processes are in the best possible shape.

Regulators realise this, and traditionally have been reluctant to allow the step up to the next level of machine learning-powered monitoring until current systems are more robust - i.e. banks improving what they’re doing today before they can consider tomorrow.

If you would like to discuss the changes your bank should be making please get in touch with Jeremy Davey or Scott Samme. In our next two blogs we’ll explain how technology and advanced analytics optimise the transaction monitoring process – and look at where new developments will take us.

Jeremy Davey|Director, Financial Crime Analytics|+44 (0)7867 502 988

Scott Samme|Director, Financial Crime Analytics|+44 (0)7753 605698

 

Comments

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been saved. Comments are moderated and will not appear until approved by the author. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment

Comments are moderated and will not appear until the author has approved them.