Probability of Default (PD) Models
Development of sophisticated PD models that accurately predict the likelihood of default across different customer segments and time horizons.
- Logistic regression and machine learning techniques
- Segmentation and portfolio analysis
- Long-run default rate calibration
- Through-the-cycle and point-in-time adjustments
- Model validation and backtesting
- Regulatory documentation and approval support