Not able to understand how the formula for calculating variance for a multi-asset equal weighted portfolio is derived as: ∑(i=1 to n) * ∑(j=1 to n) weight of i * weight of j * Cov(Ri,Rj) I am referring to page 392 on Book 6 (2014 Version)
Any insight would he very helpful. Thanks.
vsenthil:
Not able to understand how the formula for calculating variance for a multi-asset equal weighted portfolio is derived as: ∑(i=1 to n) * ∑(j=1 to n) weight of i * weight of j * Cov(Ri,Rj) I am referring to page 392 on Book 6 (2014 Version)
Any insight would he very helpful. Thanks.
You’re misinterpreting the notation; it isn’t a multiplication, it’s a _ double summation _:
∑(i =1 to n ) {∑(j =1 to n ) ([w_i_ × w_j_ × σ_i_ × σ_j_ × Cov(R_i_,R_j_)]}
This means that you let i = 1 while j runs from 1 to n , summing all the terms, then let i = 2 while j runs from 1 to n , summing all the terms, and so on, until you let i = n while j runs from 1 to n , summing all the terms; you then sum those (n ) sums. (Note, too, that you forgot the sigmas.)
S2000magician:
vsenthil:
Not able to understand how the formula for calculating variance for a multi-asset equal weighted portfolio is derived as: ∑(i=1 to n) * ∑(j=1 to n) weight of i * weight of j * Cov(Ri,Rj) I am referring to page 392 on Book 6 (2014 Version)
Any insight would he very helpful. Thanks.
You’re misinterpreting the notation; it isn’t a multiplication, it’s a _ double summation _:
∑(i =1 to n ) {∑(j =1 to n ) ([w_i_ × w_j_ × σ_i_ × σ_j_ × Cov(R_i_,R_j_)]}
This means that you let i = 1 while j runs from 1 to n , summing all the terms, then let i = 2 while j runs from 1 to n , summing all the terms, and so on, until you let i = n while j runs from 1 to n , summing all the terms; you then sum those (n ) sums. ( Note, too, that you forgot the sigmas. )
Nope, his original formula is correct. He was correct not to include the standard deviation of the two assets since the formula uses covariance.
To clarify, the correct formula is:
σp^2 = ∑(i =1 to n ) {∑(j =1 to n ) ([w_i_ × w_j_ × Cov(R_i_,R_j_)]} = ∑(i =1 to n ) {∑(j =1 to n ) ([w_i_ × w_j_ × σ_i_ × σ_j_ × correlation(R_i_,R_j_)]}
aertsommelier:
S2000magician:
vsenthil:
Not able to understand how the formula for calculating variance for a multi-asset equal weighted portfolio is derived as: ∑(i=1 to n) * ∑(j=1 to n) weight of i * weight of j * Cov(Ri,Rj) I am referring to page 392 on Book 6 (2014 Version)
Any insight would he very helpful. Thanks.
You’re misinterpreting the notation; it isn’t a multiplication, it’s a _ double summation _:
∑(i =1 to n ) {∑(j =1 to n ) ([w_i_ × w_j_ × σ_i_ × σ_j_ × Cov(R_i_,R_j_)]}
This means that you let i = 1 while j runs from 1 to n , summing all the terms, then let i = 2 while j runs from 1 to n , summing all the terms, and so on, until you let i = n while j runs from 1 to n , summing all the terms; you then sum those (n ) sums. ( Note, too, that you forgot the sigmas. )
Nope, his original formula is correct. He was correct not to include the standard deviation of the two assets since the formula uses covariance.
To clarify, the correct formula is:
σp^2 = ∑(i =1 to n ) {∑(j =1 to n ) ([w_i_ × w_j_ × Cov(R_i_,R_j_)]} = ∑(i =1 to n ) {∑(j =1 to n ) ([w_i_ × w_j_ × σ_i_ × σ_j_ × correlation(R_i_,R_j_)]}
My mistake: I was thinking correlation and typing _ covariance _. Good catch.
I’ve made the correction.