Online Program

Measuring the Distribution of Payments Fraud in a Survey of Banks: What Factors Influence Fraud Rates? (308040)

Keywords: complex survey design, fraud, nonresponse bias, missing data, imputation

Reliability of the payments system is important for a well-functioning economy, the responsibility of a central bank, and payments fraud may impose significant costs. Successful attempts at payments fraud that clear and settle are the equilibrium outcome of fraud attempts and avoidance, and in the aggregate reflect the effectiveness of payments security measures. Although cleared and settled payments fraud is a small fraction of total payments, it adds up to billions of dollars every year in the United States. To construct national estimates of fraud and rates of fraud in consumer and business payments, we survey banks to collect volumes of fraud in various payment instruments, including checks, types of electronic transfers, types of payment cards, and cash withdrawals from automated teller machines. The non-mandatory survey is a complex design, and is conducted for random samples of bank populations that are stratified by type and size. Estimation accounts for planned and unplanned patterns of missing data. We study the relationship of payments fraud to measurable factors, and examine characteristics of the distribution of fraud across banks.