Measuring and managing market, credit and Op risk [107]

Go to: Summary | Previous | Next   
Bullet points include: Suppose we have the restriction alpha = 0.01 Test using likelihood ratio test, i.e. compare following test statistic to , with, say, alpha = 0.95: Proposed by Kupiec and used by regulators when assessing accuracy of VaR models employed by banks If VaR model incorrect then we might expect that exceptions are not independent and would then need to include more general alternative, e.g. that probability of exception at time t depends on whether exception occurred at time t – 1

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile