Creating portfolio risk and return models [51]

Go to: Summary | Previous | Next   
Bullet points include: Typically assess quality of model by, e.g.: Forecasting expected VaR repeatedly through time Comparing actual outturns with predictions E.g. Basel II capital requirement depends on reliability Ideally want unconditional coverage coupled with independence Can assess coverage for entire distributional form using a quantile-quantile plot Closely related to back-testing Finding model parameters that achieve a good fit is also called calibration

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile