VaR and ES
- Thread starter txakama
- Start date
Lu Shu Kai FRM
Well-Known Member
Hi @txakama ,
I don't think there is a formula per se but rather a method of calculating. Here are the steps (anyone can create them as long as one understands the concept of ES and I assume that your 2000 observations are daily $P&L):
1. Order the losses from greatest to least (most profit).
2. Take the top 10% for 90% CI ES, which corresponds to the top 200 simulations of $P&L.
3. Take the average of these 200 simulations and that gives you the 1-day ES.
4. I am not exactly sure of this (but this can be applied to VaR), but you can scale the horizon of your 1-day ES by multiplying your 1-day ES value by the root of 21 to get your 21-day ES.
5. You can repeat steps 2 to 4 for 95% and 99% CI.
Hope this is helpful!
I don't think there is a formula per se but rather a method of calculating. Here are the steps (anyone can create them as long as one understands the concept of ES and I assume that your 2000 observations are daily $P&L):
1. Order the losses from greatest to least (most profit).
2. Take the top 10% for 90% CI ES, which corresponds to the top 200 simulations of $P&L.
3. Take the average of these 200 simulations and that gives you the 1-day ES.
4. I am not exactly sure of this (but this can be applied to VaR), but you can scale the horizon of your 1-day ES by multiplying your 1-day ES value by the root of 21 to get your 21-day ES.
5. You can repeat steps 2 to 4 for 95% and 99% CI.
Hope this is helpful!
Similar threads
