Anti-Money Laundering Transaction Monitoring Systems Implementation
Finding Anomalies
Wiley and SAS Business Series

Anti-Money Laundering Transaction Monitoring Systems Implementation provides comprehensive guidance for bank compliance and IT personnel tasked with implementing AML transaction monitoring. Written by an authority on data integration and anti-money laundering technology, this book offers both high-level discussion of transaction monitoring concepts and direct clarification of practical implementation techniques. All transaction monitoring scenarios are composed of a few common elements, and a deep understanding of these elements is the critical factor in achieving your goal; without delving into actual code, this guide provides actionable information suitable for any AML platform or solution to help you implement effective strategies and ensure regulatory compliance for your organization.

Transaction monitoring is increasingly critical to banking and business operations, and the effectiveness of any given solution is directly correlated to its implementation. This book provides clear guidance on all facets of AML transaction monitoring, from conception to implementation, to help you:

• Detect anomalies in the data
• Handle known abnormal behavior
• Comply with regulatory requirements
• Monitor transactions using various techniques

Regulators all over the world are requiring banks and other companies to institute automated systems that combat money laundering. With many variables at play on both the transaction side and the solution side of the equation, a solid understanding of AML technology and its implementation is the most critical factor in successful detection. Anti-Money Laundering Transaction Monitoring Systems Implementation is an invaluable resource for those tasked with putting these systems in place, providing clear discussion and practical implementation guidance.

About the Authors xiii

Acknowledgments xv

Preface xvii

Chapter 1 An Introduction to Anti-Money Laundering 1

The Emergence of AML 2

AML as a Compliance Domain 5

The Objectives of AML 9

Regulatory Reporting 9

Corporate Citizenship versus Profitability 10

About True and False Positives and Negatives 11

The Evolution of Automated Transaction Monitoring 15

From Rule-Based to Risk-Based 17

From Static to More Dynamic Transaction Monitoring 22

Latest Trends: Machine Learning and Artificial Intelligence 26

Latest Trends: Blockchain 29

Risk-Based Granularity and Statistical Relevance 34

Summary 36

Chapter 2 Transaction Monitoring in Different Businesses 39

Banking 43

Correspondent Banking 46

Banking – Trade Finance 49

Banking – Credit Card 60

Insurance 60

Securities 63

Stored Value Facilities (SVFs) 66

Casinos and Online Gambling 68

Lottery and Jockey Club 70

Other Businesses 72

Summary 72

Chapter 3 The Importance of Data 75

ETL: Extract, Transform, and Load 76

Extract: Data Availability and Sourcing 77

Transform: Data Quality, Conversion, and Repair 80

Data Load and Further Processing 89

Loading of the Data 89

Data Lineage 92

Multiple ETLs 92

Summary 93

Chapter 4 Typical Scenario Elements 95

Transaction Types 96

Actionable Entity 100

Scenario Parameters 106

Use of Maximum Instead of Minimum Value Threshold 108

Threshold per Customer 109

Pre-Computing Data 110

Timeliness of Alerts 112

Use of Ratios 114

Ratio as Degree of Change/Similarity 117

Ratio as Proportion 119

Other Common Issues 120

Chapter 5 Scenarios in Detail 121

Large Aggregate Value 122

Unexpected Transaction 123

High Velocity/Turnover 129

Turnaround/Round-Tripping 132

Structuring 136

Early Termination/Quick Reciprocal Action 141

Watchlist 141

Common Specifications across Unrelated Entities 142

Involving Unrelated Third Party 144

One-to-Many 144

Transacting Just below Reporting Threshold 145

Chapter 6 The Selection of Scenarios 147

Selecting Scenarios 148

Regulatory Requirements 148

Business Drivers 150

Data Quality and Availability of Reference Data 152

Maintenance of the Scenario Repository 152

How Specific should a Scenario Rule Be? 153

Overlapping Scenario Rules 155

Summary 156

Chapter 7 Entity Resolution and Watchlist Matching 157

Entity Resolution 158

Watchlists 161

Summary 184

Chapter 8 Customer Segmentation 185

The Need for Segmenting Customers 186

Approaches to Segmentation 188

Overview of Segmentation Steps 191

Organizational Profiling 193

Common Segmentation Dimensions 195

Considerations in Defining Segments 197

Check Source Data for Segmentation 199

Verify with Statistical Analysis 200

Ongoing Monitoring 205

Change of Segmentation 205

Summary 207

Chapter 9 Scenario Threshold Tuning 209

The Need for Tuning 210

Parameters and Thresholds 210

True versus False, Positive versus Negative 212

Cost 213

Adapting to the Environment 214

Relatively Simple Ways to Tune Thresholds 215

Objective of Scenario Threshold Tuning 216

Increasing Alert Productivity 216

Definition of a Productive Alert 219

Use of Thresholds in Different Kinds of Scenario Rules 220

Regulation-Driven Rules 220

Statistical Outlier 221

Insignificance Threshold 225

Safety-Blanket Rules 225

Combining Parameters 226

Steps for Threshold Tuning 228

Preparation of Analysis Data 234

Scope of Data 234

Data Columns 234

Quick and Easy Approach 237

Analysis of Dates 238

Stratified Sampling 239

Statistical Analysis of Each Tunable Scenario Threshold Variable 239

Population Distribution Table by Percentile (Ranking Analysis) 244

Distribution Diagram Compressed as a Single Line 245

Multiple Peaks 246

Zeros 246

Above-the-Line Analysis and Below-the-Line Analysis 247

Above-the-Line Analysis 247

Below-the-Line Analysis 249

Use of Scatter plots and Interactions between Parameter Variables 251

Binary Search 258

What-If Tests and Mock Investigation 260

What-If Tests 260

Sample Comparisons of What-If Tests 261

Qualifying Results of What-If Tests 262

Scenario Review Report 263

Scenario Review Approach 268

Scenario Review Results 268

Summary 274

Index 277

CHAU CHAN YIP (DEREK) is Principal Technical Consultant at SAS Hong Kong since 2010. He was formerly a Technology Consultant at Hewlett Packard. He specializes in data integration and anti-money laundering. He received his Master of Science degree in Computer Science from the Chinese University of Hong Kong.

MAARTEN VAN DIJCK NEMCSIK, LLM, PHD, has worked with SAS since 2012 as a financial crime and tax compliance domain expert and solution lead. He is part of the SAS Global Fraud & Security Business Intelligence Unit, responsible for internal and external training courses in the financial crime and compliance space.