PDA

View Full Version : #33 Historical simulation VaR


Scooterpye
05-16-2007, 01:01 PM
I noticed that a couple of people put their numerical answer for this as $3810, which I imagine is based on the percentage change in the 80th percentile event: (94.38-98.38)/98.38 = -3.81%

Seems reasonable and normal, but...

In the section on estimating volatility, Hull used the ln of the ratio of two numbers. Using this method, the change would be ln(94.38/98.38) = -3.89%, which would mean the VaR (if there's no other fault in this answer) would be $3890.

Which answer, if either, do you think is correct? Do you think the CAS gives partial credit if you used the wrong method of calculating percentage change?