Normal deviate for Operational risk VaR?

[afterworkguinness]

Hi @David Harper CFA FRM CIPM ,
The loss distribution for operational risk is not normal, it has a very heavy right tail for high severity low probability losses.

In the topic review for operational risk, we see that when calculating a VaR for operational risk, we use a normal deviate for the 99.9% confidence level. Why are we using a normal deviate when the loss distribution is not normal?

Thanks in advance.

 

[brian.field]

This is a great question!

 

[David Harper CFA FRM]

Hi @afterworkguinness sorry, where is that? (If you are referring to LVaR, it's not operational risk. The VaR is market risk and the liquidity cost is based on an assumption about the spread. But it sounds like you don't mean this ...)

 

[brian.field]

@David Harper CFA FRM CIPM - I think @afterworkguinness was referring to 302.1 C where it indicates 2.326 for a 99% z deviate (but I could be wrong).

I also think you answer the question; the LVaR is a Market Risk metric, so the normal deviate is not inappropriate.

Lastly, I find it very confusing that the Liquidity Risk topics (LVaR, etc.) are presented in the Operational Risk readings when (it seems) they are more Market Risk related.

Thanks!

Brian

 

[afterworkguinness]

I was positive I saw this, but will double check and post back. Thanks!

 

[afterworkguinness]

I went back over the notes and examples for LDA approach and the focus review and found nothing. I don't know where I came up with this notion. Thanks David and Brian.