The key theme: building next generation fincrime partnerships

Financial institutions, law enforcement agencies and regulators across the globe are realising their effectiveness in tackling financial crime is limited unless they can share information and intelligence with each other more closely.

After all, criminals have no such qualms in forming international networks themselves to traffic illicit goods, people and wildlife and launder the proceeds.

The formation of Financial Intelligence Sharing Partnerships, beginning with the United Kingdom’s Joint Money Laundering Intelligence Taskforce (JMLIT) is beginning to dominate conversation in financial crime circles.

Banks are also coming together to form private-private partnerships to improve their processes.

What should these partnerships look like? How can they be made more effective and strategic, rather than tactical?

Next month’s free global livestreaming event FinCrime World Forum features a whole series of panel debates on this very topic – building the next generation of fincrime partnerships.

To get you in the mood, we’ve picked out three partnership initiatives highlighted in a recent report by the Future of Financial Information Sharing Programme, a research partnership between the Royal United Services Institute thinktank and NJM Advisory.

Three eye-catching fincrime partnerships

Project DANIE

Set up as a pilot project in Europe to improve AML due diligence and identity verification, Project DANIE is project is run by a consortium of three investment banks and three data providers. Several other financial institutions are also involved in testing the concept.

Participants receive individual reports identifying whether there are discrepancies between their client reference data and that held by other institutions in the project.

No personally identifiable client data is shared.Privacy-Enhancing Technologies and techniques such as homomorphic encryption, secure multi-party computation and zero knowledge proof are used to ensure data privacy.

“Through the project, participants learn if other participants share a high probability match for a common client and benchmark indication of potential discrepancies in client reference data,” FFIS said.

A pilot project was launched in 2019 and it is currently being tested on real customer data.

FFIS said the project managers believe the solution is “scaleablae and could possibly be used for maket data and other AML information.

AUSTRAC Fintel Alliance Alerting Project

This project has been set up to deliver the ability for banks in Australia to identify whether there are any financial links between two (or more) suspicious accounts and the ability to trace suspicious funds as they move between accounts across financial institutions. The project uses Privacy Enhancing Technology, in the form of homomorphic encryption, to link databases held by different retail banks in the country without compromising privacy.

It is being led by regulator AUSTRAC and delivered by the public-private partnership Fintel Alliance, which consists of 28 member organisaitons.

The algorithm flags suspicious transactions from domestic retail accounts. FFIS says this closes an intelligence gap as currently domestic instructions, unlike international transfer, are not automatically reported to AUSTRAC.

The project will be tested in a simulated federated architecture using a sample of real data, before the platform is implemented in the real environment. The deployed algorithm is expected to cover more than 100 million accounts.

Duality SecurePlusTM Query for collaborative investigations in Financial Crime and Compliance

The aim of this collaboration between banks in the UK and Canada is to demonstrate the capability of banks to identify matching customer information external data sets,, without disclosing their original query to the data owner.

Two pilots are under way, one looking at fraud with four UK banks and one focused on AML involving a large Canadian bank.

FFIS said: “When receiving a query, a data owner has visibility over the query structure, but not the encrypted parameters. This ensures that the data owners know and approve the type of information they might be sharing and can control and approve specific queries execution.

“Data owners can also see an audit trail, which includes who the inquiring party was, when and how many queries were executed, and the query structure, but not the protected parameters of the queries.”

Both the fraud and AML pilots were carried out using synthetic data.

“These initiatives succeeded in validating that Homomorphic Encryption, as implemented by PALISADE and used by Duality SecurePlus Query demonstrating readiness for real-world challenges of sensitive data collaborations in financial crime investigation and compliance”, said FFIS