Skip to content
Prev 2291 / 15274 Next

Framework for VAR allocation among traders

My point is,
when underlying is non normal, any sample higher
moments may highly sensitive to outliers; without a
study of sample moments sensitity and converegence to
outliers, you can not justify the quality of VaR
modification.


you tested/simulated one skewed t distribution, but
you can not rule out all other underlying distribution
possibilities even within t-distribution with
different DOF.

These higher momonents mod on VaR are overdone IMHO.
--- elton wang <ahala2000 at yahoo.com> wrote:

            
http://www.gloriamundi.org/detailpopup.asp?ID=453055537
____________________________________________________________________________________
____________________________________________________________________________________
Be a better friend, newshound, and