Taking a Modern Approach to Fraud Prevention
Over the past 20 years financial institutions have seen an explosion in the amount and types of fraud perpetrated against them. This represents a considerable loss to business at a critical time when banks seem to be losing money hand over fist as a result of the recent sub prime mortgage crisis – in the UK alone, card fraud losses in 2006 totalled €620.6m (£428.0m).
Fraud in its many forms can also have a severely detrimental impact upon reputational risk. The unfavourable press associated with the use of card transactions to support terrorism, drugs, prostitution and human-trafficking can only result in a negative customer perception. Thus, from a risk management perspective, the requirement to actively and effectively prevent and detect fraud is absolute.
At last, in addition to the prudential requirements of effective fraud loss reduction as advised by the Joint Money Laundering Steering Group’s guidelines, another compliance driver is coming into existence. Fraud prevention and detection, as it affects credit and debit cards and other financial transactions, is incorporated in the framework of the single euro payments area (SEPA). This evolving mandate will impose additional requirements on all European financial institutions. As a result, they will need to seriously evaluate the way that they incorporate and comply to the mandate within their own risk management, fraud prevention and detection initiatives.
Neural networks (NNs) have become the standard used by many banks to detect fraud, particularly in credit cards. However it is becoming increasingly clear that such an approach does not tackle the increasingly evolving forms of fraud that banks are facing. There is a growing sentiment amongst fraud and risk managers that they are no longer comfortable trusting the score output of neural network based solutions using a model that is not representative of their card-holder base or their demographics.
If banks are to be able to detect and prevent the fast paced and evolving forms of fraud that they are facing there is a requirement for them to be able to have agile systems that are able to react in real time.
Using a system such as a business rules management system (BRMS), organisations are able to maintain all rules applied to detecting and tackling fraud in a central repository. This enables fraud and risk managers to have an overarching view of fraud prevention across all their lines of business. These rules can be easily changed and modified in business language, providing banks with the flexibility to quickly adapt to evolving conditions as criminals find new ways to commit fraud. A BRMS also delivers an improved audit trail, so banks can track the rules supporting fraud detection policies over time. This is particularly important for governance and compliance.
The success of a bank, just like any business, depends upon the quality of the decisions it makes when dealing with its customers. This is a particular problem for fraud and risk managers who are increasingly caught between the demands of reducing losses to fraud and balancing the requirements of customer value management. False positive rates (FPR) are becoming a much more significant driver of a fraud detection operation.
Essentially, a BRMS is used to validate various conditions for anomaly detection. A BRMS can support multiple attributes which can be easily modified or updated in near real-time in an effort to stay ahead of the fraudsters. Rules are used to validate millions of incoming transactions in real-time and detect anomalies based on multiple criteria, i.e. the value of the transaction; the frequency of use of the credit card: location of charges made. In doing so, BRMS offers the ability for the business to make better decisions at appropriate FPR to mitigate negative customer experience and with significantly lower implementation and running costs than a NN-solution.
The significant advances in technology combined with the attitudes to fraud prevention and detection have rendered many of the current procedures currently operating within banks obsolete. The challenge for banks is to implement systems that will be agile enough to tackle fraud in its varying forms, while also managing the customer experience. What is crucial is that banks must be able to create a safe financial transaction environment for the customer that ensures their customer-value requirements are retained without compromising the primary task of reducing losses to fraud.