In a recent meeting of the German Bundestag’s Financial Committee, Christof Schulte, announced that the Financial Intelligence Unit (FIU), which he heads, currently has no backlog of unprocessed suspicious activity reports (SARs). Despite a sharp rise in the number of SARs in 2021, Mr Schulte noted that the authority’s implementation of an innovative risk-based approach (RBA) and adoption of AI-led system, FIU Analytics, was largely responsible for successfully clearing the backlog. Here, we look at the RBA and FIU Analytics.
In 2021, the FIU received over 300,000 SARs, with more than 44,000 passed on to law enforcement agencies. To deal with such a high number of reports, the FIU applies its RBA, a global standard in processing anti-money laundering reports. The RBA is based on the idea of classifying SARs and allowing the FIU to filter by risk status before fully assessing them. This way SARs that are particularly high risk are prioritised for a more detailed review. Lower priority SARs are examined only when additional risk indicators are present. Prior to the RBA’s introduction in 2019, the FIU followed a strict ‘first in, first out’ approach to the SARs it received.
The cornerstone of the RBA are 11 predefined key risk areas for money laundering and/or terrorist financing. These key risk areas have been identified on basis of the National AML Risk Analysis (NRA) and other framework parameters, such as EU sanctions lists and evaluations of the Financial Action Task Force (a global AML watchdog). Both industry-related and phenomenon-related risks are considered. Key risk areas for money laundering include the real estate sector, trade-based money laundering and the use of cash when procuring high-value goods. For terrorist financing, the government identified a misuse of money/value transfer services NGOs/NPOs as specific risk areas. New payment methods such as cryptocurrencies constitute a risk area for both money laundering and terrorist financing.
The FIU is also required to inform enforcement authorities about indications for any other form of criminal activity besides money laundering. To fulfil this mandate, the FIU has adjusted its RBA and now also focuses on fraud cases related to COVID-19 subsidies by making this an additional (ad-hoc) key risk area.
Alongside the RBA, the FIU implemented an AI-based system called ‘FIU Analytics’. The system initially assesses the incoming data and helps FIU staff by scoring the SARs’ risk on a scale from 0 to 100. Only if the algorithm produces a high score – indicating the SAR to be related to one (or more) of the key risk areas – it is prioritised for further examination by a human investigator. Otherwise (or if the human investigator does not find sufficient risk indicators for money laundering) the SAR is added to an ‘information pool’ and automatically cross-referenced with other SARs for risk indicators. As new SARs are added to the information pool, a previously low-scoring SAR may receive a higher score and subsequently be forwarded for further review by FIU staff.
The implementation of FIU Analytics aims to combine the efficiency and accuracy of AI-based solutions with the knowledge and analytical skills of human experts. FIU Analytics is still at an early stage, having been authorised in 2019 and implemented in late 2020. According to the Ministry of Finance, it uses a ‘controlled learning environment’, indicating machine learning based on the concept of ‘supervised learning’.
A supervised learning algorithm is trained on a sample dataset (ie the already classified SARs), which has manually been classified (ie 11 risk areas plus the class of ‘no risk area’). It produces an inferred function, which can be used for mapping new examples. An optimal setup will allow for the algorithm to correctly determine the class labels for unseen SRAs.
Based on acquired skills, the AI analyses not only single SARs but also ‘big data’ (the information pool) and identifies complex correlations. This should enable the algorithm to spot money laundering schemes and patterns and help investigators to further adjust risk factors. To provide the AI with a large amount of data, the FIU was recently granted further access rights to databases of law enforcement and tax authorities.
Practical implications and outlook
With the above adjustments the FIU is now better equipped to fight money laundering and terrorist financing. As a concept of preventive measures laid down in the Geldwäschegesetz (Money Laundering Act) (GwG), which also foresees a cooperation with banks and other private sector entities. Such entities must apply a RBA themselves and forward relevant information in form of an SAR to the FIU.
To mitigate the risk of (accidently) becoming an accomplice in money laundering activities, management and reporting officers in obliged entities should consider the FIU’s key risk areas in more detail. Taking into account the authorities’ analysis could provide companies with a stronger defence argument should the adequacy of its AML risk assessment ever be tried in court. Understanding the FIU’s key risk areas will also help companies to conduct a solid assessment of when to hold a transaction longer than the three-day-period provided for in the GwG or file a complaint with the Public Prosecutor’s Office.
The innovations within the FIU and the clearance of the SAR-backlog come at a convenient time. Currently, the Financial Action Task Force (FATF) is reviewing the implementation of its standards in Germany, as well as the effectiveness of the AML framework. The results will be published in summer 2022.
However, against the backdrop of the development of the European AML framework, Germany will have to further adapt its money laundering legislation irrespective of the results of the FATF review. The EU recently issued a legislative package to overhaul its AML and counter-terrorist financing legislation. At a national and supranational level, the fight against money laundering and terrorist financing is again gathering momentum. All companies obliged under the GwG are well advised to adjust their risk analysis considering the FIU’s new approach and keep abreast of new developments. AML-related headlines are expected to continue for some time.