[R-SIG-Finance] Need help please

[Arun.stat]

hmmm..........interesting no doubt............what about this sort of
modeling?

1. 1st define a r.v. to represent the number of days/weeks price is being
kept fixed, you could try some exponential dist.
2. 2nd define another r.v. to represent the magnitude of change in prices.

Then you need to combine those two r.v., perhaps you could use "convolution"
concept (i am not sure exactly) and then simulate different price scenario
and choose the worst 5th. By this way can calculate VaR under MCS framework,
not really parametric as you wished.

However I would really be happy if someone throws points on following "

1. How to incorporate correlation (if any, if not you need to quantify still
to justify that) between above those 2 r.v.s
2. How to calculate VaR from a portfolio perspective.

I would really be happy if someone would point out the feasibility of above
approach.

Thanks,


-- 
View this message in context: http://n4.nabble.com/Need-help-please-tp1290132p1352675.html
Sent from the Rmetrics mailing list archive at Nabble.com.