Venter Financial Question

[Shouxiaozi]

I am not sure whether I did the right derivation -

my question is regarding the variance formula for AR(1) variance. In Venter paper, it is s^2*[1-b^(2t+1)]/(1-b^2); but I can only get s^2*[1-b^(2t)]/(1-b^2). Can anybody give it a try to either disprove or approve me?

Thanks.

[Colymbosathon ecplecticos]

The variance of your second term is not zero.

[Shouxiaozi]

I would think that r_i is known, so its variance is 0.

If I use the notation from Venter Financial r_t instead of r_i+t, then the second item becomes b^t * r0. Obviously here r0 is a known amount, and its variance must be zero.

Correct me if I am wrong though.

[Colymbosathon ecplecticos]

Originally Posted by Shouxiaozi

I would think that r_i is known, so its variance is 0.

True, it is known after time i, but imagine doing this at time = 0.