1. UN denotes the unconstrained model, while R I and R II are the models under Regime I and Regime II respectively. The model calibrations are specified by the training and testing windows. The results of predicted probabilities versus actual outcomes over the following 8Q testing period are used to calculate the loss metric for 90+ days delinquencies within 8Q;Notes: Performance comparison of GBT models of consumer default risk under various monotonicity constraint regimes,2016
2. Do Banks Pass Through Credit Expansions to Consumers Who Want to Borrow?;Sumit Agarwal;The Quarterly Journal of Economics,2014
3. Credit Growth and the Financial Crisis: A New Narrative
4. Machine learning methods that economists should know about;Susan Athey;Annual Review of Economics,2019