We have considerable experience in developing models for bank portfolios where the lack of default data poses a critical constraint on the use of traditional statistical modeling techniques. We, along with our global partners, have developed a suite of techniques specifically focused for low default scenarios.
- Concordance maximization with expert rankings for PD models under Low Default Portfolio (LDP) scenarios
- Bayesian Framework based methods for LGD and EAD models under low loss and recovery information scenarios
- Bespoke modeling techniques for Zero Default Portfolios (ZDP), which include methods derived from Case Based Reasoning