Highlights
Data sourcing, ingestion and transformation
Calibration to map into model inputs, model enhancements
Enhanced, in-depth profit & loss assessment
Balance sheet and capital optimisation
Business appetite assessment and opportunity identification
Situation
Riskcare has developed a high-performance, multi-factor Monte Carlo simulation engine and methodology for the calculation of Default and Voluntary Termination (VT) losses for large portfolios of UK Retail Loans.
Action / Approach
Introduced additional risk factors and new data inputs – data sourcing and normalisation, assessment of individual variables for their quality as predictors and calibration work to map data into model inputs.
Rewrote and enhanced existing methodology - incorporating new data as well as enhancements to exposure modelling, addition of loans according to amortisation target in each month of the revolving period and simulation of new loans’ events/losses.
Enhanced existing code to optimise performance - added stress testing and calibration capabilities, user GUI for incorporating SME assumptions to be applied to forecasts, “what-if” stress or business scenarios.
Thoroughly tested, documented and successfully supported model validation (internal and regulatory).
Result(s)
The bank was able to:
- Carry out more enhanced, more in-depth risk analyses for Significant Risk Transfer (SRT) securitisation transactions and related RWA reduction exercises in a greatly reduced amount of time.
- Run a large number of “what-if” scenarios for regulators as well as to assess business lending appetite.
- Model entire balance sheet, with easy process of adding new products and new risk factors (where applicable).