Agenda

Agenda

Risk Model Validation: Agenda

Live virtual course | Agenda timing is in GMT/EST

Start: 2pm GMT / 9am EST
End: 5.15pm GMT / 12.15am EST

14:0015:30

The origin of risk models

02:00 - 03:00

  • A short history of risk
  • A (first) definition of risk
  • The uses and misuses of statistics in risk modelling
  • Limits of statistical risk models
  • Machine learning in risk modelling?

15:3015:45

Break

03:00 - 03:15

15:4517:15

Elements of risk models – and risk model failures

03:15 - 04:15

  • Typical setup of a quantitative risk model
  • How can a risk model fail?
  • Design and implementation
  • Processes
  • How should a risk model be used?

Start: 2pm GMT / 9am EST
End: 5.15pm GMT / 12.15am EST

14:0015:30

Building a roadmap for validation

02:00 - 03:00

  • What is validation?
  • The ‘when’ – risk model validation in different stages
  • The ‘who’ and the ‘how’ – roles and expectations
  • General rules or risk model validation
  • What do regulators think about risk model validation?

15:3015:45

Break

03:00 - 03:15

15:4517:15

Toolbox 1: machine learning / market risk

03:15 - 04:15

  • Validation of model results by statistical methods
  • Examples from market risk – signatures of model failure
  • The role of learning and adaption in financial markets

Start: 2pm GMT / 9am EST
End: 5.45pm GMT / 12.45am EST

14:0015:00

Toolbox 2: credit portfolio models

02:00 - 03:00

  • The regulatory framework: Targeted Review of Internal Models (TRIM)
  • The dangers of small probabilities
  • Benchmarking – the use of alternative risk models
  • The merits and dangers of simplified risk models
  • Scenario analysis

15:0015:15

Break

03:00 - 03:15

15:1516:45

Toolbox 3: credit risk

03:15 - 04:15

  • Statistical methods for validation of data 
  • Case study: validation of risk model parameters
  • Methods for software validation 

16:4517:00

Break

04:15 - 04:30

17:0017:45

Looking back and looking ahead

04:30 - 05:30

  • Model risk governance and model inventory
  • What about interconnected risk models?
  • What do consumers of risk model results want?
  • Conclusion: risk model frameworks