Risk-Based Payment Fraud Detection

Payment frauds are a form of intrusion where money is transferred from a victim’s bank account to the bank account of a fraudster. This paper reviews why in practice, it is neither possible nor economical to prevent 100 percent of payment frauds. It therefore becomes important to detect payment frauds and to control their cost. The paper then combines two unique insights into the nature of payment fraud to propose a new risk-based payment fraud detection method. This method does not try to detect individual fraudulent payments but rather seeks to quantify the expected loss from frauds over a given time period. This expected
loss is the risk posed by frauds, and fraud managers only need to intervene if this risk exceeds a maximum acceptable loss threshold. Below this threshold, payment fraud related losses represent a contained risk, which should be viewed as an operating cost just like shoplifting is an operating cost in retail. The paper critically appraises the risk-based method and discusses its applicability in practice.