On its surface, de-risking banking activities is a justifiable move. Rather than attempt to manage the risk, banks choose instead to avoid risk. With that move, they sidestep past onerous regulator penalties, reputation-tarnishing headlines, and angry shareholders and customers.
The result, of course, is predictable. Banks find themselves in the clear but watch as revenue streams close. In their place, payment service providers (PSP) are filling the void, facilitating transactions across borders to all corners of the globe.
An Alternative Approach to De-Risking
For years, banks have tried to manage risk through rule-based tools. It didn’t really work. ThetaRay’s Chief Customer Officer Idan Keret recently wrote that rule-based approaches triggered so many false alarms that only 1% of all alerts turned out to be true positives.
With that type of success rate, it’s easy to understand why banks felt it was better to de-risk than to manage risk.
However, technology has advanced far beyond rules. Introducing artificial intelligence and machine learning tools transaction monitoring has the potential to shift the equation in correspondent banking.
Banks can and should be looking at managing risk, and taking back their cross-border payment business.
What Does it Mean to Manage Risk?
There are essentially two types of risk that banks need to manage. Once those risks are brought under control, banks can pursue correspondent banking business, expanding their relationships with respondent banks and reopening their revenue stream.
Compliance risk management is one of the bigger risks that they face. In the first six months of this year, regulators handed out over $660 million in fines and penalties for having inadequate money laundering and terrorist financing controls in place. These banks’ crimes and weren’t found to have committed any wrongdoing; they were simply assessed fines for their lack of control.
In contrast, operational risk management pushes banks to do a better job screening and monitoring transactions, as well as prevent suspicious transactions from being processed. Failing to do so puts them at risk of significant fines.
Introducing advanced detection tools that utilize artificial intelligence and machine learning can allay both types of risk. Since 2018, regulators have realized that innovations in AI and ML enable improvements in AML activities that are not achievable in any other way. That year, Dr. Lael Brainard told delegates at a conference that was optimistic about the potential for AI and ML in the fight against money laundering.
Brainard, who has sat on the Federal Reserve Board of Governors since 2014, noted that AI had superior pattern recognizability, offered cost efficiencies for banks, and was better than conventional approaches at working with large, less structured data sets.
By using AI tools to monitor transactions, banks can manage, rather than de-risk, correspondent banking.
Benefits of AI in Risk Management
Utilizing AI enables banks to get back into the correspondent banking business, and with it, reopen this revenue stream. With an effective AI tool in place to monitor transactions, banks do find themselves with additional growth opportunities.
Every year since 2012, the Basel Institute of Governance has released its annual Basel AML Index. This index ranks countries in terms of their overall risk level. They assess the risks of money laundering and terrorist financing for 141 countries.
The 2020 index showed that the average risk score, out of ten, is 5.22, and their index shows that numerous countries are vulnerable to money laundering, terrorist financing, and other related crimes.
Under a rule-based monitoring system, banks would look at the Basel AML Index or some similar tool, and use the data to build their rules. They may decide to reject all transactions that originated or terminated in any country that scored a 7 or higher, or they might combine it with a different rule, stating that any transaction that was valued at over $10,000 and originated in a country that scored 6 or higher would be rejected.
The rules lacked any opportunity for nuance, and any transaction that met its criteria would be dealt with in the manner that the rules defined.
AI changes that entire equation. Rather than following broad rules, AI tools look at entire data sets, searching for anomalies within the transaction. When it determines that a financial transfer that originated in Norway and is heading to Yemen, it can scour the details of the transaction, and determine whether that transaction can pass through on its own merits.
By taking a case-by-case approach using highly scalable AI tools, banks can process more transactions, and increase their revenues, without taking on additional compliance or operational risk.
When banks de-risk, they lose out on revenues and open the door to their competition to step in. By managing their risk with AI tools, they have the opportunity to expand their correspondent banking activities while growing revenues.