资源描述
Financial Stability Institute FSI Insights on policy implementation No 18 Suptech applications for anti-money laundering by Rodrigo Coelho, Marco De Simoni and Jermy Prenio August 2019 JEL classification: C45, G21, G38, O32 Keywords: anti-money laundering, suptech, innovation, data analytics FSI Insights are written by members of the Financial Stability Institute (FSI) of the Bank for International Settlements (BIS), often in collaboration with staff from supervisory agencies and central banks. The papers aim to contribute to international discussions on a range of contemporary regulatory and supervisory policy issues and implementation challenges faced by financial sector authorities. The views expressed in them are solely those of the authors and do not necessarily reflect those of the BIS or the Basel-based committees. Authorised by the Chairman of the FSI, Fernando Restoy. This publication is available on the BIS website (bis). To contact the BIS Media and Public Relations team, please email pressbis. You can sign up for email alerts at bis/emailalerts.htm. Bank for International Settlements 2019. All rights reserved. Brief excerpts may be reproduced or translated provided the source is stated. ISSN 2522-2481 (print) ISBN 978-92-9259-287-5 (print) ISSN 2522-249X (online) ISBN 978-92-9259-288-2 (online) Suptech applications for anti-money laundering iii Contents Executive summary . 1 Section 1 Introduction . 3 Section 2 Data-related challenges faced by AML/CFT authorities . 6 Section 3 Data analytics tools . 7 Tools developed or used by AML/CFT supervisors . 8 Tools developed or used by financial intelligence units . 10 How do AML/CFT authorities develop these tools? . 12 Issues associated with the development and use of these tools . 13 Section 4 Conclusions . 15 References . 17 Glossary . 18 Suptech applications for anti-money laundering 1 Suptech applications for anti-money laundering1Executive summary Use by financial authorities of advanced data collection and analytics tools enabled by new technologies is collectively called suptech. In the area of data analytics, development of such tools has been facilitated by advances in artificial intelligence and its practical application in machine learning, natural language processing and other advanced analytics capabilities. These tools have provided opportunities to enhance financial authorities capacity. Detecting potential anti-money laundering (AML) and combating the financing of terrorism (CFT) violations is one field where data analytics tools seem more advanced. This paper therefore dives deeper into these tools. In particular, it aims to explore the various data analytics tools used by authorities tasked with AML/CFT responsibilities, as well as their practical experiences in using such tools. Nine AML/CFT authorities are covered in this paper. Such authorities have either supervision or financial intelligence functions, or both. AML/CFT supervision and financial intelligence functions have different mandates. Authorities with AML supervision functions are expected to ensure compliance by financial institutions with requirements to combat money laundering (ML) and terrorist financing (TF). Authorities with financial intelligence functions, ie financial intelligence units (FIUs), meanwhile, are expected to serve as national centres for the receipt and analysis of suspicious transaction reports and other information relevant to money laundering, and to disseminate the results of that analysis. FIUs sometimes also have AML supervision functions. Both AML/CFT supervisors and FIUs need advanced data analytics tools to analyse the large volumes of information at their disposal. AML/CFT authorities typically receive substantial amounts of transactional and non-transactional data. On top of these traditional sources of data, some AML/CFT authorities are now actively collaborating with other government agencies and private entities to expand the scope of data available to them. Some authorities are also exploring the use of non-traditional sources of information (eg newspaper articles, social media) and integrating them with traditional information to come up with richer analyses. The difference in mandates does not seem to affect the types of advanced data analytics tools the AML/CFT authorities are pursuing. AML/CFT authorities covered in the paper are in general pursuing similar advanced data analytics tools, such as network analysis, natural language processing, text mining and machine learning. These tools increase their ability to detect networks of related transactions, to identify unusual behaviours and in general to transform significant amounts of structured and unstructured data into useful information that contributes to their respective processes. Authorities have used different strategies to develop these tools. AML/CFT authorities that are within the central bank, or prudential or conduct authority, generally benefit from the institutional 1Rodrigo Coelho (rodrigo.coelhobis) and Jermy Prenio (jermy.preniobis), Bank for International Settlements, and Marco De Simoni (marco.desimonibancaditalia.it), Bank of Italy. The authors are grateful to the representatives from the authorities interviewed; to the participants of the second FSI meeting on the use of innovative technology in financial supervision held in Basel on 5-6 June 2019 for the insightful discussions; and to Kuntay Celik, Matei Dohotaru, Tom Neylan and Greg Sutton for helpful comments and insights. We are also grateful to Esther Knzi and Cissy Mak for valuable support with this paper. 2 Suptech applications for anti-money laundering strategy to utilise innovative technology to help in supervision work and can develop these solutions in-house. For some AML/CFT authorities, taking advantage of ready solutions in the market may be more efficient. Others are actively collaborating with the academic community and promoting research in this field. Many of the authorities use a combination of these approaches. The optimal solution for a specific authority will depend on several factors such as the profile of the authority, the characteristics of the financial system that the authority oversees, and the legal framework in which the authority operates. Efficiency gains seem to be the number one benefit of advanced data analytics tools, which could help capacity-constrained AML/CFT authorities. AML/CFT authorities have highlighted the gains in terms of time savings they achieved in using these advanced data analytics tools. This could translate to reallocation of resources or capacity from more manual work to more judgment-based work. Assessing effectiveness particularly of tools used by FIUs, however, is not that straightforward. The benefits that these tools bring are particularly important for jurisdictions that have been heavily impacted by the unintended consequences of AML/CFT international standards, particularly de-risking. Jurisdictions most frequently exited by global correspondent banks seem to be those with weak AML/CFT supervisory and regulatory frameworks. That being so, the development of advanced data analytics tools as well as development of necessary skills could help strengthen these jurisdictions frameworks and potentially reverse this trend. However, the use of these innovative technologies gives rise to a number of challenges. First, computational capacity may be an issue, since these tools deal with large volumes of data. Second, data privacy and confidentiality requirements provide safeguards that AML/CFT authorities must consider in using certain data and external resources in developing data analytics tools. Third, assessing the effectiveness of these tools might be challenging, in particular for FIUs given the necessary time to prove the occurrence of a money laundering activity. Finally, tools based on supervised machine learning could lose their effectiveness over time, especially if not regularly updated with new training data, given the capacity of criminal organisations to change their behaviour in order to avoid detection. There is scope for information-sharing among AML/CFT authorities on the data analytics tools they are developing or using in order to promote peer learning. Although the data analytics tools used by AML/CFT authorities are tweaked to reflect their mandates, the underlying methodologies of these tools are quite similar. There are therefore opportunities for peer learning through regular exchange of information and sharing of experiences on the development and use of these tools. AML/CFT authorities that are just starting to develop their data infrastructure have a “late mover” advantage and may find it easier to integrate advanced data analytics tools. These authorities have the advantage of developing their data infrastructure from scratch without the burden of legacy systems. They can design it in a way that makes the data collection, validation and management processes seamless, while more easily enabling the integration of newly developed analytical tools. ML/TF risks have international reach, so development of data analytics tools that are international in scope should be considered. The tools discussed in this paper are all national in scope. Money laundering, however, is an international issue, and criminal organisations tend to exploit loopholes anywhere in the world. Therefore, a strong argument could be made for international cooperation and collaboration in terms of developing data analytics tools with an international coverage. Suptech applications for anti-money laundering 3 Section 1 Introduction 1. Use by financial authorities of advanced data collection and analytics tools enabled by new technologies is collectively called suptech.2In the area of data analytics, development of such tools has been facilitated by advances in artificial intelligence (AI) and its practical application in machine learning, natural language processing (NLP) and other advanced analytics capabilities. These tools have provided opportunities to enhance financial authorities capacity. 2. These suptech data analytics tools can be used by authorities for anti-money laundering (AML) and combating the financing of terrorism (CFT) purposes. Broeders and Prenio (2018) provides a comprehensive overview of the suptech tools used by different authorities. In the area of data analytics, suptech tools can be found in market surveillance, misconduct analysis, microprudential supervision and macroprudential supervision. Detecting potential AML/CFT violations is one field where data analytics tools for misconduct analysis are used and seems more advanced. 3. The use of these tools has the potential to significantly contribute to reducing money laundering (ML) and terrorist financing (TF) risks by strengthening jurisdictions AML/CFT frameworks. One estimate of AML/CFT compliance cost by financial firms put it at USD 25.3 billion per year for US financial services firms alone.3AML/CFT authorities are also devoting substantial resources to fight ML/TF to the extent possible. However, many of them, particularly those in emerging market and developing economies, have severe capacity constraints. Despite these efforts, estimates of money laundered worldwide are still staggering.4In addition, there seems to be slow progress globally in reducing ML/TF risks.5This has wide-ranging economic and social repercussions. The Financial Stability Board (FSB), for example, noted that jurisdictions most frequently exited by global correspondent banks seem to be those with weak AML/CFT supervisory and regulatory frameworks.6This has implications for international trade and remittance flows to these jurisdictions. 4. In particular, these tools can help AML/CFT authorities7analyse the large volumes of information at their disposal. AML/CFT authorities typically receive substantial amounts of transactional data from reporting institutions. These are then supplemented by non-transactional data (eg data from tax, customs and property registration authorities) to provide context. Some AML/CFT authorities are now actively collaborating with other government agencies and private entities to expand the scope of data available to them. Some authorities are also exploring the use of non-traditional sources of information (eg newspaper articles, social media) and integrating them with traditional information to come up with richer analyses. The emergence of big data analytics enables the integration of all this information from different sources to tell a coherent story. 5. Advanced data analytics tools could also potentially improve the effectiveness of AML/CFT authorities. Financial institutions traditionally relied on rule-based AML/CFT measures, which are shown to generate false positive alerts of around 9095% leading to substantial resource implications.8So they 2Broeders and Prenio (2018) defines suptech as the use of innovative technology by supervisory agencies to support supervision. For this paper, we broadened the definition of suptech to include advanced data collection and analytics tools used by financial authorities, ie not only supervisors or regulators but other authorities as well, such as financial intelligence units (FIUs). 3LexisNexis Risk Solutions (2018). 4See for example MONEYVAL (2017). 5See eg Basel Institute of Governance (2018). 6FSB (2015). 7In this paper, the expression “AML/CFT authorities” is meant to enco
展开阅读全文