Research On Transaction-Based Suspicious Money Laundering Behaviroal Patterns And Anti-Money Laundering Countermeasures

Money laundering (ML) is publicly recognized as one of the "non-traditional security issues" after the Cold War threatening financial stability, social fairness, government reputation, public order, human ethnics, and social harmony. It is a rational decision to engage in ML activities, but money laundering itself is not a rational behavior in traditional terms, implying that money launderers and their activities are characteristically different from the normal economic behaviors of normal people. So it is of a great theoretical significance as well as of a promising social practicability for this dissertation to make a research on some basic theoretical issues regarding money laundering and anti-money laundering (AML), the suspicious behavioral patterns of trade-based money laundering (TBML) and account-based money laundering (ABML) on the basis of transactions, and the corresponding AML countermeasures.The definition of money laundering is theoretically extended in the dissertation from traditional "conversion of black money into white money" to include "conversion of black money into grey money" and "conversion of grey money into white money" and the nature of money laundering is explored from an inter-discipline perspective. It argues that the AML research in China can be divided into three periods and typical events are identified for each accordingly. The continnum theory of transactional patterns and the four cardinal principles of national AML undertaking are expounded in theory.The research pioneers a systematic analysis of TBML both at home and abroad. TBML refers to the apparently legitimate ownership of the property shifted by falsifying trade truth. The "transfer pricing-money laundering-capital flight interrelationship model" is formulated first. Then, it constructs pricing strategy (portfolio) models and gives ML scale estimate methods under transfer pricing-based money laundering (TPBML). It goes further to discuss capital flight, income tax evasion, and import duty evasion as a result of money laundering in case of imports at inflated prices and exports at deflated prices in general trade. From the discussions of a transit trader's imports at both inflated and deflated prices followed by exports at inflated/deflated prices, respectively, in transit businesses, that is, inflated import-inflated export, inflated import-deflated export, deflated import-inflated export, deflated import-deflated export, it particularly recognizes money laundering under the disguise of deficits on account surface. It finds that the selection of banking system in money laundering by means of remittance, collection, and letters of credit (L/Cs) is determined by the credit foundations of payment terms themselves. It also finds that money laundering co-exists with trade and flies to wherever international trade extends, which severely endangers the world trading system. It is concluded that TBML is most likely to become the most popular ML method due to the increasing AML efforts in international finance industry, exploitation of international jurisdictional issues, and the imperfection of international jurisdictional assistance, etc. Blueprints are thus presented as to the establishment of a multi-perspective anti-TBML framework.ABML aims to illegally transfer and take charge of property through a series of operations in transactional amount, transactional frequency, transactional interval, transactional partner, and transactional manner, etc. in bank accounts correspondence. The research argues to distinguish accidental money laundering from habitual money laundering for relative studies. Based on the categorization of ABML research perspectives and the continuum theory of transactional patterns, the research constructs and experimentally testifies suspicious ML transactional behavioral pattern recognition algorithms from both diachronic and synchronic points of view after extracting and describing subjective features and behavioral features of suspicious ABML activities. It is found that taking transaction amount, transaction amount deviation coefficiency, withdrawal frequency, and deposit frequency as variables, distance-based clustering and clustering-based local outlier detection (improved CBLOF algorithm) can be effectively applied to discover accidental ML behavioral patterns from historical sequential data mining of one single account. As well, habitual ML behavioral patterns can be effectively detected from synchronic comparative analysis of peer group accounts by means of improved grid clustering algorithm and density-based local outlier factor algorithm, with transaction amount deviation coefficiency, transaction frequency, cash/withdrawal ratio, and cash/deposit ratio being research variables. It is helpful for the AML transition from feedback control to early prevention and in-process surveillance to build a real-time, dynamic, and self-adaptable recognition system of suspicious transactions.Based on the above analyses, AML countermeasures are discussed from the perspectives of laws and regulations, national commitment, and financial institutions' (FIs') risk-based approach. A special attention is paid to the exploration of the endogenous assumptions of the current AML regime in institutional economics and social control terms as they have not deterred money launderers as expected. It argues that AML regime should guarantee FIs' change from defensive over-reporting under fear of fines for non-compliance into voluntary filing of informative SARs under drive of benefits and promote FIs' integrating SAR submission with operational risk management. A nation should devote to establish regulative AML systems and pan-society AML systems for the sake of a total process control of money laundering activities. Guided by the notion of risk segmentation and control, FIs should effect appropriate Know-Your-Customer (KYC) or due diligence procedures and transaction monitoring procedures in conformity with ML risk categories and levels.Nevertheless, money laundering is a dynamic self-learning organization system in nature, national AML undertaking should therefore seek for the balance between the effectiveness of regulations and the efficiency corrupted by those regulations.