Journal of Credit Risk

Risk.net

Forecasting consumer credit recovery failure: classification approaches

Hyeongjun Kim, Hoon Cho and Doojin Ryu

  • We analyze actual data on credit recovery program applicants and find key determinants of credit recovery failure.
  • We present forecasting models to predict credit recovery failure.
  • Artificial neural network algorithms can make better predictions than logistic regressions or other methodologies can.

This study proposes an advanced credit evaluation method for nonperforming consumer loans, which may serve as a new investment opportunity in the post-pandemic era. Our results, based on both a unique account-level data set and machine learning techniques, imply that the artificial neural network algorithm with demographic and account-related variables performs the best in terms of predicting consumer credit recovery failure within 24 months. We also find that the key determinants of such failures are the total amount of delinquent debt, the applicant’s age and the maximum length of the overdue period. A forecasting model using the random forest algorithm can also be improved by using additional information that is determined after a debtor applies for the credit recovery program. Our findings have practical implications for banks, financial institutions and investors who need to manage and evaluate nonperforming loans.

To continue reading...

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here: