The document discusses using R for credit risk modeling tasks at a bank. It provides examples of using R to model mortgage haircut rates, calibrate through-the-cycle (TTC) credit risk models, and bin customers into cohorts for TTC probability of default estimation. R is shown to provide more flexibility than SAS for tasks like fitting customized statistical models with varying distributions and volatility, conducting TTC calibration via backcasting, and specifying customer cohorting.