Nematrian
/

Stress Testing and Back-testing for UCITS [21]

Go to: Summary | Previous | Next   
Bullet points include: Commercial risk systems often assume that future returns will be (log)-Normally distributed, and therefore may not cope well with extreme events, i.e. ‘fat tails’. But these are precisely the ones that generate the most press. Even if models do assume non-Normality, how much should we assume? Some risks like liquidity risk are very fat-tailed and very difficult to model . Hence, increasing focus on stress testing across all regulatory frameworks. But how different is it in reality to VaR and other ‘likelihood’ based approaches?

NAVIGATION LINKS
Contents | Prev | Next | Library

Desktop view | Switch to Mobile