https://www.polity.org.za
Deepening Democracy through Access to Information
Home / Recommendations RSS ← Back
Close

Email this article

separate emails by commas, maximum limit of 4 addresses

Sponsored by

Close

Article Enquiry

Fixing AML: Can New Technology Help Address the De-risking Dilemma?

Close

Embed Video

Fixing AML: Can New Technology Help Address the De-risking Dilemma?

Fixing AML: Can New Technology Help Address the De-risking Dilemma?

25th April 2018

ARTICLE ENQUIRY      SAVE THIS ARTICLE      EMAIL THIS ARTICLE

Font size: -+

  • Fixing AML: Can New Technology Help Address the De-risking Dilemma?
    Download
    12.40 MB
Sponsored by

In 2015, a Center for Global Development (CGD) working group on the unintended consequences of anti-money laundering (AML) and countering the financing of terrorism (CFT) policies argued that the policies that have been put in place to counter financial crimes may also have unintentional and costly consequences for people in poor countries. The report’s authors identified the problem of “de-risking.” In other words, AML/CFT policies have had a chilling effect on banks’ willingness to undertake cross-border transactions. Banks are unwilling to do business in markets perceived to be risky (or low in transaction volume) in part because of the perceived high cost of compliance.

Who are the losers from de-risking? The analysis of the CGD working group points to the families of migrant workers; small businesses that need to access working capital or trade finance; and recipients of lifesaving aid in active conflict, post-conflict, or post-disaster situations. And sometimes, AML/CFT policies may be self-defeating to the extent that they reduce the transparency of financial flows.

Advertisement

One consequence of de-risking may be the decline in correspondent banking services, which are critical to cross-border financial transactions. Worldwide, the number of correspondent banking relationships has declined by more than 6 percent since 2011. Small and fragile countries have been especially affected.

Even while policy solutions to address de-risking are being implemented, new technologies have emerged to address de-risking by increasing the efficiency and effectiveness of AML/CFT compliance by financial institutions. These new technologies may enhance transparency and information-sharing capabilities, facilitate automation and interoperability between platforms and institutions, mutualize certain compliance functions, and improve banks’ ability to accurately identify illicit activity.

Advertisement

This report, to the best of our knowledge, is the first comprehensive effort to assess six key new technologies and their potential to solve the de-risking problem. These include know-your-customer (KYC) utilities, big data, machine learning, distributed ledger technology (DLT), legal entity identifiers (LEIs), and biometrics. These new technologies (in use and on the horizon) may make it easier to conduct AML/CFT compliance, which in turn might tip banks’ cost-benefit calculation and make holding correspondent banking accounts with clients in poor countries more likely.

With a view to educating policymakers, regulators, and the broader audience interested in addressing the de-risking problem, we describe what these technologies are and how they work. We examine what parts of the AML/CFT compliance workflow they can improve, including customer identification and verification, customer due diligence, and transaction monitoring. We also examine the limitations of these technologies and the barriers to adoption they face. Finally, we offer recommendations for how policymakers and regulators can responsibly support the adoption of these technologies.

KYC utilities are central repositories for customer due diligence (CDD) information. By centralizing information collection and verification, KYC utilities can reduce the amount of information that has to be exchanged bilaterally between correspondent banks and their respondents, thereby reducing the time banks spend conducting CDD investigations. KYC utilities may also help facilitate the adoption of a baseline dataset for CDD information. Several KYC utilities were launched in 2014 and 2015, catering to different client segments, including the correspondent banking sector. In one business model, the information provider (in the case of correspondent banking, the respondent bank) uploads its information for free, and the information consumer (the correspondent bank) pays a fee to access the information. Although KYC utilities emerged organically in response to market demand, most industry bodies have urged regulators to produce more explicit guidance as to how much banks may rely on these services, in order to further increase utilization.

Big data refers to datasets that are high in volume, high in velocity, and high in variety, and therefore require systems and analytical techniques that differ from those used for traditional datasets. Compared with relational databases, big data applications offer more scalable storage capacity and processing. They also allow many different types of data to be stored in one place, so compliance staff spend less time gathering information from disparate sources. Most important, they can greatly expand the range and scope of information available for KYC and suspicious transaction investigations. Big data applications are typically paired with advanced analytics engines—including machine learning programs—that can help identify complex patterns and relationships in the data that might have otherwise gone undetected.

Machine learning is a type of artificial intelligence—itself a branch of computer science—that allows computers to improve their performance at a task through repeated iterations. There are three broad types of machine learning—supervised, unsupervised, and reinforcement learning. With supervised learning, the machine learning program analyzes a dataset to build a model that best predicts a predefined output. In contrast, with unsupervised learning, the machine learning program is not given a predefined output—rather, it explores the data on its own, looking for patterns and relationships in the dataset. Reinforcement learning falls between the other two, with the algorithm receiving general feedback on its performance, but without a specific predefined output to aim for. Machine learning may be used to augment or transform a number of compliance functions, including those for developing more sophisticated customer typologies and for more accurately monitoring transactions. These uses could simultaneously cut down on false alerts and identify new or hitherto undetected illicit finance techniques. Banks may benefit from more leeway to explore these new technologies. Banks would also benefit from more government feedback on the suspicious activity reports (SARs) they file, which would help them to further hone their detection capabilities.

DLT is a way of securely organizing data on a peer-to-peer network of computers. In a blockchain, which is a type of DLT, data modifications, such as transactions, are recorded in time-stamped blocks. Each block is connected to previous blocks, forming a chain. Modifications are confirmed and stored by all users on the network, which makes the ledger difficult to tamper with. Although blockchain technology is most commonly associated with virtual currencies, such as Bitcoin, the basic technology has a number of other potential use cases, including uses in regulatory compliance. In particular, DLT may be used for securely storing and sharing KYC information, as well as for cheaper and more secure international payments. This technology is yet to be widely adopted, but single-use cases are emerging in different parts of the world.

LEIs are unique alphanumeric identifiers, like barcodes, that connect to reference datasets held in a public database. Any legal entity that makes financial transactions or enters into contracts may request an LEI. In many countries, especially developed ones, LEIs are increasingly mandated by regulation. To date, more than 1 million LEIs have been issued worldwide. By serving as common identifiers, LEIs can enable different platforms, organizational units, and institutions to refer to entities clearly and without any ambiguity. This interoperability can, in turn, facilitate greater automation and information sharing. In addition, the reference datasets can serve as a starting point for CDD. A further extension of the LEI would be to include it in payment messages to identify originators and beneficiaries, which would further enhance the transparency of international payments. However, this would require changes to payment message formats and to banks’ IT systems, as well as more widespread adoption of the LEI outside of the financial sector and also in developing countries.

Biometrics use distinctive physiological or behavioral characteristics to authenticate a person’s identity and control his or her access to a system. Natural persons are not eligible for LEIs except in limited circumstances, so a separate standard is needed for identifying individuals. Biometrics are more robust than other authentication factors, such as passwords and tokens, as they are generally more secure and easier to use. Biometrics are being used to address the “identification gap” that exists in many developing countries. This use, in turn, could make it easier for banks to conduct customer identification, verification, and due diligence, which may bolster the confidence of their correspondent banks. However, most biometric identification systems are being developed at the national level, meaning that individual identification is still fragmented at the international level, hindering the use of biometrics for international payments. Work is required to develop an internationally recognized and interoperable identification system for natural persons.

Report by the Centre for Global Development

EMAIL THIS ARTICLE      SAVE THIS ARTICLE ARTICLE ENQUIRY

To subscribe email subscriptions@creamermedia.co.za or click here
To advertise email advertising@creamermedia.co.za or click here

Comment Guidelines

About

Polity.org.za is a product of Creamer Media.
www.creamermedia.co.za

Other Creamer Media Products include:
Engineering News
Mining Weekly
Research Channel Africa

Read more

Subscriptions

We offer a variety of subscriptions to our Magazine, Website, PDF Reports and our photo library.

Subscriptions are available via the Creamer Media Store.

View store

Advertise

Advertising on Polity.org.za is an effective way to build and consolidate a company's profile among clients and prospective clients. Email advertising@creamermedia.co.za

View options
Free daily email newsletter Register Now