Page 1 of 1
A stat question
Posted: August 16th, 2004, 12:15 pm
by SPAAGG
Hi, assume I have 2 random variables, X and Y, correlated. I observe N realizations from both, and can compute x.hat and y.hat, the sample means from x.i and y.i observed.Could you give me an approximation formula for Var(x.hat /y.hat)I have heard about the delta method. Could someone explain me !tx David
A stat question
Posted: August 16th, 2004, 12:15 pm
by SPAAGG
Var is the variance of course, not the Value at Risk...
A stat question
Posted: August 17th, 2004, 8:14 pm
by battant
hi,for the variance of the portfolio, the response is : sum of the variance + 2*Covariancefor the delta method, it's used , under given conditions , for calculate the Value at Risk. ANother method is delta gamma method