How money laundering episodes catalyze the battle against financial crime

Money laundering currently ranks as one of the world’s most prolific activities, with an astonishing $2 trillion laundered annually—yet, a staggering 99% of these illicit transactions go undetected. In recent years, numerous high-profile cases have been uncovered in Australia, Malaysia, and most recently, Singapore. 

Given that these locations are considered financial hubs, these revelations sparked questions as to how their reputations would be impacted. The very fact that the fraudulent transactions were uncovered, followed by swift, resolute actions against those responsible, underscores the commitment to safeguarding their status as trusted financial centers. 

It is imperative that countries join forces with global regulators to turn these adversities into a canvas for change—an opportunity to fortify controls and redraw the framework for collaborative monitoring. 

Building on its current presidency of the intergovernmental Financial Action Task Force and its role as a sandbox for financial innovation, Singapore has an opportunity to revolutionize global due diligence practices by building the future of mutualized compliance using distributed-ledger technology (DLT). 

Today, anti-money laundering and counter-terrorism financing (AML/CTF) monitoring at an entity level has witnessed significant advancements, harnessing the power of cutting-edge AI technologies such as fuzzy logic and network link analysis. Fuzzy logic addresses the shortcomings of rule-based systems, while network link analysis uncovers complex relationships. 

These innovations have enabled organizations to paint a more precise and comprehensive picture of their clients’ profiles and activities within the confines of their own operations. Nevertheless, these advancements still fall short against increasingly sophisticated evasive tactics of bad actors. 

For instance, acts of funds diversion involving the rapid transfer of nominal sums across different accounts, assets, and intermediaries are prone to evade detection at each step of the process. It is a stark reality that underscores the importance of global collaboration and adherence to recognized industry standards. 

Take, for instance, the Wolfsberg Principles developed by 13 banks, which have since served as the cornerstone of compliance programs of banks worldwide. Without the concerted effort of regulators and financial institutions around the world, perpetrators can easily exploit vulnerabilities in the financial system. 

In unity, there is strength 

More work needs to be done at the systemic level. A potential global solution is mutualized compliance that requires the standardization and integration of know your customer (KYC) and transaction monitoring systems across financial institutions. This approach, akin to the creation of an “Interpol of the digital world”, seeks to enable financial institutions to collaborate and share financial intelligence on a global scale without compromising the confidentiality of client data. 

One way to achieve this is through the creation of a global utility comprising decentralized digital identities integrated with data repositories feeding real-time changes in a banking user’s background and federated machine learning that enables distributed training without data exchange. 

Decentralized digital identities are the user’s truest and latest identities in real time, put together by regulated financial institutions, which the Monetary Authority of Singapore defines as “trust anchors”. Digital identities contain a collection of verifiable credentials (VCs) that represent user characteristics without revealing personally identifiable information through privacy-preserving tools like zero-knowledge proof. 

For example, a trust anchor can issue a VC attesting that “Tom” is above 18 years old upon presentation of his driver’s license, all without revealing his actual age. Given the dynamicity of user characteristics, it is crucial that expired VCs—issued VCs that are no longer applicable to the user—can be revoked by trust anchors to ensure that the digital identity is updated in real time against data repositories. 

As trust anchors collectively contribute to the wealth of information encapsulated by digital identities, users can leverage them to instantaneously access financial services around the globe. Analogous to a “create once, reuse many times” KYC, this eliminates duplicated efforts in supplying and validating the same KYC information at each financial institution. 

Not only does this streamline the client onboarding process for both financial institutions and their clients, but it also creates a holistic, immutable, and auditable record of activities the user has engaged in across entities and jurisdictions. Leveraging these comprehensive records, deeper connections and insights can be drawn to flag suspicious persons or activities more effectively. 

As Aristotle observed, the whole often transcends the sum of its constituent parts. In the realm of financial oversight, this principle finds resonance in collaborative monitoring. It is key to unlocking insights that would remain elusive if financial institutions were to monitor in isolation. 

Numerous central banks, including Singapore’s, have embarked on the development of information-sharing platforms. These platforms enable participating financial institutions to exchange critical information regarding suspicious clients or transactions. This approach marks a significant milestone for the industry as it seeks to depart from the conventional, siloed method of transaction monitoring, which has proven highly inadequate. 

However, a crucial concern looms large on data privacy. These initiatives entail the sharing of sensitive proprietary data that could not only create regulatory repercussions but also erode competitive advantage of financial institutions. As such, they encounter the formidable challenge of securing industry adoption that is essential to generate the scale and network effect pivotal to their effectiveness.

This is where federated machine learning is set to redefine the problem. As a decentralized form of machine learning, models are trained locally by individual financial institutions using data residing within their own independent nodes. 

These data points are created each time a client uses their digital identity to engage in transactions with the financial institution. To achieve data uniformity and reliability, central banks can play the pivotal role of developing and distributing standardized machine learning algorithms and input variables to financial institutions. 

This algorithm moves from one financial institution to another to train on their data, all the while preserving the sanctity of data security as no data is transferred in the process (see Figure 1). The result is the creation of a robust model cultivated through an expanded dataset, spanning diverse volumes and real-world scenarios across the financial system. 

This greatly enhances financial institutions’ abilities to swiftly reject suspicious transactions and improve true positive detection rates. Such an approach also has the added benefit of converging AML/CTF monitoring processes globally to a golden standard. 

At the intersection of reinvention

Through the innovative implementation of digital identity management systems like SingPass (for individuals) and CorpPass (for companies), Singapore has pioneered a transformative paradigm for the secure digital storage and utilization of personal information. We believe Singapore is poised to extend its expertise in the formation of a global digital identity framework where multiple stakeholders work in mutuality to preserve the safety and integrity of our financial system. 

Jie Yi (Jaelynn) Lee is a graduate analyst at Deutsche Bank. Mathew Kathayanat is head of product, securities services for Asia-Pacific at Deutsche Bank.