/


Creating and validating risk models [42]

Go to: Summary | Previous | Next   
Bullet points include: Risk measurement should primarily be about the future, not the past Therefore should be attractive to incorporate implied volatility or other contemporaneous data in risk model formulation ‘Fully’ market consistent risk measurement focuses on the market prices of risk exposures, will use market-implied data where possible Even when we do not believe that market-implied data will provide a good guide to future ‘real world’ outcomes, see e.g. Kemp (2009)  or Kemp (2005) Back-testing such models is more difficult Need to strip out impact of time-varying risk premia Not in historical returns but present in ‘fully’ market consistent measures

NAVIGATION LINKS
Contents | Prev | Next | ERM Lecture Series


Desktop view | Switch to Mobile