Fraud Prevention and Detection for Credit and Debit Card Transactions

Transcription

1 August 2009 Fraud Prevention and Detection for Credit and Debit Card Transactions Richard Collard Senior Business Lead and SME - Market Development ILOG, Software Sales, IBM Sales and Distribution

2 Page 2 Contents Preface 2 Fraud Loss: A Cost of Doing Business? 2 Credit and Debit Card Fraud 3 Neural Networks: Time for Change? 4 Card Fraud Detection Using Business Rules 5 BRMS: Flexibility, Empowerment and Ubiquity 5 Preface Over the past 20 years, financial institutions, governments, insurers and retailers have seen an explosion in the amount and types of fraud perpetrated against them. In the United Kingdom alone, card-fraud losses in 2006 totaled million ( 428 million) and while this total represented a reduction of 3 percent over 2004 and a decrease of nearly 116 million ( 80 million) over the past two years, it was still a considerable loss to business. Of particular concern is the evolution of types of fraud to circumvent the effectiveness of PIN-based domestic transactions. This has led to a 43 percent increase in fraud committed on UK cards abroad, where perpetrators take advantage of non-pin environments. Fraud Loss: A Cost of Doing Business? BRMS and Card Fraud Detection: The Way Forward 7 UK banks recently reported their total profits for 2006 amounted to 40 billion. Considering the size of this figure, it can be said that fighting card fraud is not wholly related to financial loss but rather to a significant risk to the banks reputations. The negative press associated with the use of fraudulent card transactions to support terrorism, drugs, prostitution and human trafficking can only result in a negative customer perception. Therefore, from a risk-management perspective, it is important to actively and effectively prevent and detect card fraud. The growth of organized crime and terrorism and their associated requirements are well documented. Their need for significant funding easily explains the inventiveness and increasing sophistication of criminal gangs and individuals in their attempts to defraud organizations of huge sums on a global scale. The manifestations of fraud are seen in money laundering,

3 Page 3 ID theft, internal/collusive fraud, threats to homeland security, account takeover, transactional fraud on card and checking accounts -- the list goes on and proves that countering fraud effectively requires a fast response with a multilayered approach. Fraud prevention and detection, as it affects credit and debit cards and other financial transactions, is incorporated into the framework of the Single European Payments Area (SEPA). In addition to the prudential requirements of effective fraud loss reduction, an additional compliance driver is coming into existence. Fraud prevention and detection, as it affects credit and debit cards and other financial transactions, is incorporated into the framework of the Single European Payments Area (SEPA). This evolving mandate will undoubtedly impose additional requirements on all European financial institutions. As a result, financial institutions will need to examine how they address this subject. Credit and Debit Card Fraud Over the last 15 years, the card industry has tended to espouse neural network (NN)-based solutions as the de facto standard for preventing and detecting fraud. Given the prevalence of such systems and the significant associated outlay in terms of license fees and implementation costs, it has been difficult for providers of alternative systems to mount a case against incumbent NN solutions. Furthermore, the mystique woven around blackbox solutions has contributed to condemning the alternatives to the periphery. But this is no longer the case. NN has drawbacks that are becoming too significant to ignore and warrant reconsideration of more effective alternatives.

4 Page 4 The value of IBM WebSphere ILOG Business Rule Management System (BRMS) approach to fraud detection can no longer be ignored given the significant changes to the technology and associated business attitudes over the last few years. Neural Networks: Time for Change? The value of a business rule management system (BRMS) approach to fraud detection can no longer be ignored given the significant changes to the technology and associated business attitudes over the last few years. Fraud and risk managers increasingly agree that they can no longer trust the scores given by a NN-based solution that uses a model which does not accurately represent their card-holder base or their demographics. Customer retention and loyalty are greatly impacted by negative experiences such as wrong account debits. This has led to a growing number of commercial decisions that compromise prudential risk management. The provision of viable false positive rates (FPR) to achieve this end is a prerequisite and clean FPR mappings can be created in a BRMS. The threat posed by fraud to the reputations of financial institutions is considered a key driver in the implementation of effective fraud prevention and detection solutions. The relatively high cost and limited availability of appropriate geographically and demographically relevant NN models is no longer acceptable. NN models cannot be adapted quickly to significant behavioral changes such as what happened when chip and PIN cards were introduced. NN-based systems rely on rules to detect frauds that are not covered by the models used by the systems (for example, flash frauds and evolving scams). Transposition of an NN model from credit card fraud detection to a debit card or an anti moneylaundering environment is not feasible.

5 Page 5 A major Global 500 financial institution has been very successful using a BRMS to detect fraud. Card Fraud Detection Using Business Rules A major Global 500 financial institution has been very successful using a BRMS to detect fraud. Business rules are used to validate various conditions for detecting anomalies that can indicate fraud. The performance of the rules is high enough to provide real-time detection of anomalies based on several criteria, including multiple sources, transaction values, card-use frequency, merchant and location of the charges. BRMS provides a user-friendly point and click environment that helps business users to create and modify fraud detection rules offline. Rules can be created, modified and tested quickly and then deployed to a production system when ready. This enables institutions to react quickly in their effort to keep pace with fraudsters. New detection policies can be activated in hours, instead of months, helping to reduce lost revenue and increase customer satisfaction. BRMS: Flexibility, Empowerment and Ubiquity The key value of adopting BRMS to provide fraud detection capabilities lies in the flexibility that this methodology offers from an installation and a business-use perspective. BRMS offers the ability to use a common platform to address fraud issues throughout an organization, removing the need to identify different solutions and platforms to tackle credit card, debit card, check and money-laundering fraud.

6 Page 6 Downstream, operational functions in the transactional process, i.e. chargebacks, dispute resolution, authorizations, risk and loyalty can all be managed effectively by a BRMS. BRMS can effectively manage downstream, operational functions in the transactional process, i.e. chargebacks, dispute resolution, authorizations, risk and loyalty. The fraud and risk manager is increasingly caught between demands to reduce losses to fraud and balance the requirements of customer value management and marketing. With FPR becoming a significant driver for fraud detection operations, it is vital that a solution offer realistic FPR values to help the fraud prevention and detection strategy to be updated, tailored and adopted quickly and its performance to be monitored efficiently. BRMS offers businesses the ability to make better decisions based on FPR to mitigate customers negative experiences and with significantly lower implementation and running costs than an NN solution, a compelling argument exists for reevaluating the systems deployed today. A BRMS-based fraud prevention and detection solution affords: Superior transparency and understanding of the processes In-house control of detection targets Significant flexibility to respond quickly to evolving fraud types A simpler migration path to help parallel existence with incumbent solutions Business users complete control of strategic and tactical decisions The ability to implement rules that reflect risk and commercial business drivers

7 Page 7 In essence, a BRMS applied to the task of card fraud prevention and detection helps credit and risk personnel to interact effectively in both strategic and tactical efforts with the commercial and customer-value requirements of their organization without compromising the primary task of reducing losses to fraud. BRMS and Card Fraud Detection: The Way Forward Advances in BRMS technology strongly suggest that BRMS-based solutions provide a viable way forward to ensure that the fundamental goals of fraud prevention and customer satisfaction are achieved as effectively as incumbent solutions but with a significantly lower cost of ownership. Any non-brms-based fraud prevention and detection solution that is commercially available will use rules to supplement the core detection paradigm. Advances in BRMS technology strongly suggest that BRMS-based solutions provide a viable way forward to ensure that the fundamental goals of fraud prevention and customer satisfaction are achieved as effectively as incumbent solutions but with a significantly lower cost of ownership. To IBM ILOG, significant changes in technology and attitudes toward fraud prevention and detection for credit and debit cards help to present the BRMS approach as an extremely viable ally in the fight against card fraud. To find out more, please visit

8 Page 8 Copyright IBM Corporation 2009 IBM Corporation Software Group Route 100 Somers, NY U.S.A. Produced in the United States of America August 2009 IBM, the IBM logo, ibm.com and WebSphere are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol ( or ), these symbols indicate U.S. registered or common law trademarks in other countries. A current list of IBM trademarks is available on the Web at Copyright and trademark information at ibm.com/legal/copytrade.shtml Other product, company or service names may be trademarks or service marks of others. IBM assumes no responsibility regarding the accuracy of the information provided herein and use of such information is at the recipient s own risk. Information herein may be changed or updated without notice. IBM may also make improvements and/ or changes in the products and/or the programs described herein at any time without notice. References in this publication to IBM products and services do not imply that IBM intends to make them available in all countries in which IBM operates. Recyclable, please recycle. WSW14053-USEN-01