Historical VaR

[[email protected]]

The schweser notes say the 99% 300 days VaR using historical simulation is the 3rd worst result.

However, one of the 2010 GARP practice questions used the 4th worst result.

My practical experience also tells me it should be the 4th one.

What do u think?

 

[David Harper CFA FRM]

Hi Vincent, many forum threads on this issue. In practice, there is not correct/incorrect answer: 3rd, 4th, and interpolations, too, between 3rd and 4th are all valid.

We follow the FRM assignment (Dowd) who suggests (defaults to, without conviction) the 4th worst, like you;
this because you can then identify the 3 worst as the 1% tail (3/300 = 1%) and say "we expect the 1% to be WORSE than the 4th" rather than "we expect the 1% to be at least as bad as the 3rd." I find this just a little bit more natural!

(the inconvenience is that Jorion actually uses the 3rd
... we've asked GARP to pick one just for the sake of the exam. Still waiting)

David