We recently wrote about the issue of banks de-risking by exiting relationships with customers in high risk or exiting entire lines of business. Here at QuantaVerse, we oppose de-risking for several reasons such as:
- Leads to loss of valuable revenue streams;
- Invites regulatory scrutiny by abandoning jurisdictions and business lines, and;
- Creates the perception that the de-risking bank has inadequate risk mitigation tools to distinguish between honest customers and criminal ventures.
Now, the Office of the Comptroller of the Currency (OCC) has issued guidance directing banks to adhere to a set of best practices prior to deciding to eliminate foreign correspondent banking relationships, which are as follows:
- Establish effective governance for overseeing how banks reevaluate risk and monitor recommendations for retaining or terminating foreign correspondent accounts;
- Regular communication to senior management about decisions to retain or terminate foreign correspondent accounts while considering any adverse impact that closures may have on access to financial services for an entire group of customers or an entire region;
- Establish communications with foreign correspondent customers in the context of determining whether to withdraw from a relationship;
- Consider specific information these customers may provide that might mitigate risks they present;
- When decisions are made to terminate accounts, provide sufficient time for customers to establish alternative banking relationships, unless any delay would create additional risk, and;
- Ensure a clear audit trail of the reasons and method used for account closure.
By effectively leveraging their own data sources through a solid data science solution, banks can then make informed and pragmatic de-risking decisions for foreign correspondent banks and the ultimate beneficial owner/customer. A data science approach is highly automated and efficient, and it allows banks to retain revenue without significantly driving up AML and KYC costs associated with that revenue.