Suppose that you calculate VaR from Duration. In your many results you ?nd that:
i. using historical data (of whatever length) or a normal distribution does not a?ect the result; 11
ii. you ?nd that kurtosis between historical data and the normal distri- bution is almost identical;
iii. You ?nd the expected change in the portfolio ?P = 0, with very small standard errors. Given the above, can you say that this Duration based VaR is an appro- priate approach to measure risk?

Question

Suppose that you calculate VaR from Duration. In your many results you ?nd that:
 i. using historical data (of whatever length) or a normal distribution does not a?ect the result; 11
 ii. you ?nd that kurtosis between historical data and the normal distri- bution is almost identical; 
iii. You ?nd the expected change in the portfolio ?P = 0, with very small standard errors. Given the above, can you say that this Duration based VaR is an appro- priate approach to measure risk?

Quizplus · Accepted Answer

The Answer of Suppose that you calculate VaR from Duration. In your many results you ?nd that:
 i. using historical data (of whatever length) or a normal distribution does not a?ect the result; 11
 ii. you ?nd that kurtosis between historical data and the normal distri- bution is almost identical; 
iii. You ?nd the expected change in the portfolio ?P = 0, with very small standard errors. Given the above, can you say that this Duration based VaR is an appro- priate approach to measure risk?

Suppose That You Calculate VaR from Duration