Advanced ERM Session 8: Formulating, implementing and validating risk models
This presentation is based on a part of an academic course on Advanced Enterprise Risk Management (Advanced ERM) titled ‘Formulating, implementing and validating risk models’ and covers topics such as: model selection (including methodologies and parameterisation), model validation, evaluating inputs / implementation and evaluating outputs
Slides
| 1 | Session 8: Formulating, implementing and validating risk models |
| 2 | Session 8: Formulating, Implementing and validating risk models |
| 3 | Model selection: general issues |
| 4 | Example users and uses of models |
| 5 | Models may: |
| 6 | Session 8: Formulating, Implementing and validating risk models |
| 7 | Example: credit risk models and methodologies |
| 8 | There are different dimensions to risk |
| 9 | Credit risk with and without interest rate risk |
| 10 | Stochastic spreads also potentially important |
| 11 | Risk versus pricing |
| 12 | How might we create a general framework? |
| 13 | Short-term volatility models |
| 14 | Longer holding period models |
| 15 | Fundamental data limitations |
| 16 | Session 8: Formulating, Implementing and validating risk models |
| 17 | Introduction to parameterisation |
| 18 | Notation |
| 19 | Factor structure |
| 20 | Example factor structure (1) |
| 21 | Example factor structure (2) |
| 22 | Estimating factor loadings, e.g. |
| 23 | How might we calculate pair-wise correlations? |
| 24 | Benchmarking |
| 25 | Correlation structure implicit in Basel II |
| 26 | Session 8: Formulating, Implementing and validating risk models |
| 27 | Validation |
| 28 | E.g. a credit risk model |
| 29 | E.g. internal ratings assessments |
| 30 | Possible approaches include |
| 31 | Possible performance measure, for default models |
| 32 | Validation of statistical inputs |
| 33 | Validation of implementation |
| 34 | Session 8: Formulating, Implementing and validating risk models |
| 35 | Evaluating outputs |
| 36 | Model risk and back-testing |
| 37 | Model assessment |
| 38 | Link with calibration |
| 39 | Ideally model should also fit well 'period by period' |
| 40 | Look-back bias and in-sample testing |
| 41 | Look-back bias and out-of-sample testing |
| 42 | Market consistent risk measurement |
| 43 | Back testing in other regulatory frameworks |
| 44 | Session 8: Agenda covered |
| 45 | Important Information |
NAVIGATION LINKS
Contents | Next | ERM Lecture Series