Module 1: Introduction to Model Risk
(No exercises)
Module 2: Model Risk Management
(No exercises)
Module 3: Model Risk Management Guidelines – SR 11-7 (US)
(No exercises)
Module 4: Guide for the Targeted Review of Internal Models (TRIM – EU)
(No exercises)
Module 5: Quantitative Measurement of Model Risk
(No exercises)
Module 6: Models to Quantify Model Risk
Exercise 1: Estimation of a model risk scorecard, definition of scale and ranges.
Exercise 2: Estimation of model risk using frequency and severity distributions (Poisson, Negative Binomial; Lognormal, Burr, Gamma, Weibull, Inverse Gaussian, GDP EVT, 4-parameter g-and-h, lognormal mixtures, lognormal–EVT, Poisson–Gamma Bayesian, lognormal–GDP partition, expert-based scenarios).
Exercise 3: Selection of the best distribution using goodness-of-fit tests (Cramér–von Mises, Anderson–Darling, Kolmogorov–Smirnov).
Exercise 4: Comparison of model risk using Panjer recursion, Fast Fourier Transform and Monte Carlo simulation.
Module 7: Advanced Data Validation
Exercise 5: Exploratory data analysis in R.
Exercise 6: Detection and treatment of outliers in R.
Exercise 7: Missing value imputation techniques in R.
Exercise 8: Stratified and random sampling.
Exercise 9: Weight of Evidence (WOE) analysis in Excel.
Exercise 10: Univariate percentile analysis in R.
Exercise 11: Optimal univariate analysis for continuous variables in R.
Exercise 12: KS, Gini and Information Value validation for each variable in R and Excel.
Exercise 14: Optimization of categorical variables in R.
Exercise 15: Univariate analysis with decision trees in R.
Exercise 16: Segmentation using K-means clustering in R.
Module 8: Multivariate Models and Machine Learning
Exercise 17: Logistic regression with stepwise selection in R.
Exercise 18: Piecewise regression in Excel and R.
Exercise 19: Decision trees in R.
Exercise 20: Support Vector Machine in R.
Exercise 21: Neural networks in R.
Exercise 22: Ensemble models in R.
Exercise 23: Random Forest in R.
Exercise 24: Bagging in R.
Exercise 26: Comparison of discriminant power between neural networks and logistic regression models.
Exercise 27: Model risk using confidence intervals of logistic regression coefficients.
Module 9: Model Risk in the Scorecard
Exercise 28: Construction of a scorecard in Excel.
Exercise 29: Optimal cut-off point estimation in Excel and model risk from cut-off selection.
Exercise 30: Confusion matrix to verify Type I and Type II error in Excel with and without variables.
Module 10: Stability Tests
Exercise 31: Stability tests for models and factors.
Module 11: Discriminant Power of Traditional and Machine Learning Models
Exercise 32: Estimation of Gini, Information Value, Brier Score, Lift curve, CAP, ROC and divergence in SAS and Excel.
Exercise 33: Jackknifing.
Exercise 34: Bootstrapping.
Exercise 35: Kappa statistic estimation.
Exercise 36: K-fold cross-validation in R.
Exercise 37: Traffic-light (semaphore) validation out-of-time (6-year horizon) for logistic and ML models.
Module 12: Model Risk in Credit Scoring
Exercise 38: Application of a credit scoring tool over a 2-year horizon without updating data.
Exercise 39: Measurement of model risk via Type I/II errors and losses from accepting bads and opportunity cost from rejecting goods.
Module 14: Explainable Artificial Intelligence (XAI)
Exercise 40: XAI-based interpretability of credit scoring.
Module 15: Automation of Modelling
Exercise 41: Automation of credit scoring modelling, hyperparameter optimization and validation.
Module 16: PD Calibration and Backtesting
Exercise 42: PD calibration using central tendency adjustment.
Exercise 43: PD calibration using quasi-moment matching.
Exercise 44: PD calibration using regression approaches.
Exercise 45: PD calibration in machine learning models.
Exercise 46: PD backtesting in Excel.
Exercise 47: Forecasting PD and realized PD in Excel.
Exercise 48: Validation using Monte Carlo simulation.
Module 17: Alternative PD Backtesting
Exercise 49: PD validation using a latent-variable model.
Module 18: LGD Calibration and Backtesting
Exercise 50: Model performance comparison using calibration and accuracy tests.
Module 19: EAD Backtesting
Exercise 51: Performance comparison of EAD models.
Module 20: Lifetime PD Estimation and Backtesting (IFRS 9)
Exercise 52: Machine learning to estimate Lifetime PD.
Exercise 53: Multinomial regression to estimate Lifetime PD.
Exercise 54: Cox regression to estimate Lifetime PD.
Exercise 55: Multi-stage Markov chains in R.
Exercise 56: Stability and accuracy testing of Lifetime PD.
Exercise 57: Lifetime PD calibration for machine learning models.
Module 21: Estimation and Validation of Expected Credit Losses (ECL) – IFRS 9
Global Exercise 58: Lifetime ECL estimation (Lifetime PD, LGD, EAD, prepayment probability and scenarios).
Global Exercise 59: Validation through ECL simulation and economic capital.
Module 22: IFRS 9 ECL Validation
(No exercises)
Module 23: Economic Capital Models
Exercise 60: Default correlation in retail/consumer portfolios.
Exercise 61: Asset correlation with EMV (e.g., Exogenous Maturity Vintage).
Exercise 62: Sectoral CreditRisk+ implementation.
Exercise 63: Single-factor model.
Exercise 64: Multi-factor model.
Exercise 65: Copulas in R.
Module 24: Validation and Model Risk in Economic Capital
Exercise 66: Capital validation using the Berkowitz test.
Exercise 67: Model risk due to uncertainty in CreditRisk+.
Exercise 68: Model risk due to data deficiencies in CreditRisk+.
Module 25: Model Risk – Parameter Uncertainty in Credit Risk
Exercise 69: Model risk measurement in an ASRF credit risk model (parameter uncertainty in credit scoring, PD and LGD calibration).
Exercise 70: Model risk of VaR and ES under parametric lognormal distribution.
Exercise 71: Model risk from PD and asset-correlation parameter uncertainty using a Bayesian approach.
Exercise 72: Propagation of uncertainty in credit risk models.
Module 26: Statistical Validation of Models
Exercise 73: Comprehensive analysis of regression statistical tests.
Module 27: Estimation and Validation of Financial & Macroeconomic Time Series Models
Exercise 74: Non-stationary and cointegrated series analysis.
Exercise 75: GARCH modelling of market variables.
Exercise 76: Machine learning modelling (SVM and neural networks).
Exercise 77: Backtesting of machine learning time series models.
Module 28: Determination of Macroeconomic Scenarios in IFRS 9
Exercise 78: Advanced macroeconomic scenario modelling with BVaR and DSGE.
Module 29: Stress Testing Net Charge-Off
Exercise 79: VAR stress testing model.
Exercise 80: VEC stress testing model.
Exercise 81: MARS stress testing model.
Module 30: Stress Testing Validation
Exercise 82: Validation tests comparing VAR vs MARS stress testing models.
Module 31: Model Risk in IFRS 9 Stress Testing
Exercise 83: Model risk in stress testing using the Credit Portfolio View approach.
Exercise 84: Model risk in the Multiyear approach.
Exercise 85: Model risk in LGD stress testing.
Exercise 86: Model risk in transition matrix stress testing.
Module 32: Margin of Conservatism (MoC) – IRB
Exercise 87: Estimation of MoC for PD, LGD and EAD under two different methodologies.
