Linking Real-Time Information to Actions: Collectability Scores for Delinquent Credit-Card Accounts

In August 2009, the Federal Reserve Bank reports the volume of consumer credit-card debt in the United States to be in excess of $900 billion. According to the 2009 Nilson Report this number is projected to grow by 20% in 2010. Developing an optimal strategy for collecting such an enormous debt is a crucial operational problem that, to the best of our knowledge, has not been successfully studied in either academia or industry. We provide a new approach to the estimation of the repayment probability based on account-specific and macroeconomic information. Our model generates a probability measure of the collectability of an account balance. Unlike the FICO score, which is a relative index of creditworthiness of an individual and does not have any intrinsic meaning, our collectability score corresponds to the actual chance of collecting a given percentage of an account debt over a given time horizon. Furthermore, in contrast with FICO score, our collectability score is specialized for overdue credit-card accounts placed in collection. ∗Ph.D. Candidate, Department of Management Science and Engineering, Terman Engineering Center, Stanford University, Stanford, CA 94305-4026. Email: [email protected] †Assistant Professor, Department of Management Science and Engineering, Terman Engineering Center, Stanford University, Stanford, CA 94305-4026. Phone: (650) 725-6827. Email: [email protected].