VaR Daily vs Annual

H

hlmasterchief

New member
Joined
Jun 18, 2026
Messages
0
Reaction score
0
There’s an example in Kaplan secret sauce and just wondering what am I doing wrong.
Daily Std Dev = 0.0011
1Day Rtn = 0.00085
Value of Portfolio=100m
So var at 5% significance:
VaR(daily) = [0.00085 - 1.65(0.0011)] * 100m = -96,500
I am trying to convert this to Annual VaR now considering 250 days
VaR(annual) = [0.00085*250 - 1.65(0.0011*(sqrt(250)))]* 100m = 18,380,233
Why my VaR is positive?
Thanks in advance for any help.
 
What is the right answer according Kaplan ?
 
Paste this in google
0.00085*250-1.65*(0.0011*15.8)*100000000
-2,867,699.78
 
This is due to MV multiplication at the end. Take that out and var is positive. :/
 
Google this
0.00085*250-1.65*(0.0011*15.8)
or
[0.00085*250-1.65*(0.0011*15.8)]*100000000
 
Oh, now I see it.
The Sharpe ratio for annual VAR is 12.07.
Makes sense since the daily VAR is 0.09% of the portfolio, and that scales down when you increase the time period.
 
nah, it’s just the figures are a little bit unfortunate here - the daily var was slightly negative but because you multiply returns by 250 and volatilty by only 250^0.5, the outcome happens to be positive
 
hlmasterchief wrote:So nothing I am doing wrong :)
Click to expand...
Well … .
I ran a Monte Carlo simulation using the data you gave (assuming that the returns are normally distributed), compounding the daily return for 250 days. The results are:
  • Mean annual return = 2.37% (which is consistent with 0.85% compounded for 250 days)
  • Annual standard deviation of returns = 2.15%, which is a lot more than 0.11% × √ 250 (=1.7393%)
That’s probably why they didn’t compute the statistics for the annual return: it isn’t remotely as simple as you’d imagine.
 
MrSmart wrote:I don’t think compounding mean returns is a good idea.
Click to expand...
Maybe not.
But the data aren’t clear on whether that mean return is to be compounded or not.
 
S2000magician wrote:
MrSmart wrote:I don’t think compounding mean returns is a good idea.
Click to expand...
Maybe not.
But the data aren’t clear on whether that mean return is to be compounded or not.
Click to expand...
I mean from a statistical prespective, since this is a single period model, and we’re using standard deviation.
 
You must log in or register to reply here.
Back
Top