Analysis of extreme events [42]

Go to: Summary | Previous | Next   
Bullet points include: Requires estimation of substantially increased numbers of parameters Usually from left tail, so fewer data points available from which to estimate, making estimates less reliable Scherer (2007) suggests ‘best practice’ involves Specify candidate distributional form type Find best fit to data from within this candidate type Calculate lpm’s (and thence optimal portfolios etc) from best fit Fitted cubic approach described earlier is an example of such an approach!

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile