# Derivation of arithmetic variation of a portfolio over multiple periods [closed]

I am very confused on how to derive the attached equation (15).

Would someone be kind enough to walk me through the proof?

• Use the fact that $\mathbb{V}[r_{t,t+k}] = \mathbb{E}[r_{t,t+k}^2] - \mathbb{E}[r_{t,t+k}]^2$ and use equations (6), (12) and (13) to compute these expectations. – JejeBelfort Oct 6 '17 at 7:16
• Hello, Thank you so much for your response @JejeBelfort!! If you expand the PI term from n=1 to k, and subtract 1. You are left with a long equation of r terms multiplied with each other which you then square. how do you simplify this proof and see the pattern? also where does the V(of R_t) come from in the answer? E(r^2)-E(r)^2 from the expansion I mentioned? – jungsun65813215 Oct 6 '17 at 7:50
• Since the $r_{t+n}$ are uncorrelated, you can write $\mathbb{E}\left[\prod_{n=1}^k r_{t+n}\right] = \prod_{n=1}^k \mathbb{E}\left[ r_{t+n}\right]$ – JejeBelfort Oct 6 '17 at 11:31
• Why delete the attachment? The problem was: if the one day returns are iid $\mu,\sigma^2$, what is the variance of k-day returns? The original attachment is here imgur.com/a/t5OqN – noob2 Oct 6 '17 at 21:53
• @molly Don't vandalize your own question. – Bob Jansen Oct 7 '17 at 13:59

Now, since $r_{t+n}$ are independent random variables we have