Nematrian
/

Stress Testing and Back Testing [30]

Go to: Summary | Previous | Next   
Bullet points include: Particularly useful for portfolios which are less diversified, non-linear or where underlying risks are highly skewed. E.g. bond / CDS return distributions highly skewed if issuer defaults, pegged currency values can jump dramatically if peg comes unstuck. More ‘subjective’ than VaR and potentially more time consuming. How do you decide which scenarios/stresses to test and how extreme they should be? Past extremes may not reflect the future? Some commentators would like to ‘reinsert’ likelihoods into stress testing. E.g. Brewer (2009) adopting ‘meta’ probability distributions

NAVIGATION LINKS
Contents | Prev | Next | Library

Desktop view | Switch to Mobile