0
$\begingroup$

The variance on the difference between two random variables can be calculated as: Var(X-Y) = sigma^2_X - sigma^2-Y - 2 Cov(X,Y). Numerically, i got the following values:

sigma^2_X = 0.03 (3%)

sigma^2_Y = 0.05 (5%) Cov(X,Y) = 0.00025

Calculation:

Var(X-Y) = 0.03^2 + 0.05^2 - 2 * 0.00025 = 0.0029

However, I get another result if I multiply by 100 %

3^2 + 5^2 - 2 * 0.00025 = 33.9995

How do I calculate this correctly?

$\endgroup$
  • $\begingroup$ "$Y = 40 \% , X = 60 \%$". What is the meaning of that ? $\endgroup$ – callculus Oct 19 '16 at 16:51
  • $\begingroup$ I deleted it. I wrote it on accident. $\endgroup$ – DBE7 Oct 19 '16 at 16:53
  • $\begingroup$ And what next ? $\endgroup$ – callculus Oct 19 '16 at 17:13
  • $\begingroup$ Notice that the square of $3\%$ is not $9\%$. $\endgroup$ – Michael Hoppe Oct 19 '16 at 18:17
0
$\begingroup$

It is true that $Var(X-Y)=Var(X)+Var(Y)-2cov(X,Y)$.

Thus $Var(X-Y)=0.03+0.02-2\cdot 0.00025=0.0495$

There is no need to multiply anything by 100. Also 0.02 and 0.03 musn´t be squared, because they are already the values of the variances.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.