0
$\begingroup$

How do you calculate the correlation of an asset to a portfolio, when for all assets in the portfolio you know there: correlation to each other, volatility and weight in portfolio.

For example: Assets 1,2,3&4 all have volatility of 15%. Assets 1&2 have a correlation of 1 and all other pairs of assets correlation = 0.

With a portfolio of 16.7% in Assets 1 & 2 and 33.3% in 3 & 4, What I am reading states that all assets (1,2,3,4) have a correlation of 0.577 with the portfolio.

How is this calculated? Is there a formula that can be applied to broader examples with more varied asset volatilities and correlations?

$\endgroup$
3
$\begingroup$

This is to basic for this website, but I will answer it anyway as I think it is interesting.

You have a correlation matrix of 4 assets (1, 2, 3, 4). This is how it looks:

$ Correl = \begin{bmatrix} 1 &1& 0& 0 \\ 1 &1& 0& 0 \\ 0 &0& 1& 0 \\ 0 &0& 0& 1 \\ \end{bmatrix}$

Thus the covariance matrix is (show this yourself):

$ Cov= \begin{bmatrix} 0.0225 &0.0225& 0& 0 \\ 0.0225 &0.0225& 0& 0 \\ 0 &0& 0.0225& 0 \\ 0 &0& 0& 0.0225 \\ \end{bmatrix}$

The weights matrix is:

$w= \begin{bmatrix} 0.167 \\ 0.167 \\ 0.333 \\ 0.333 \\ \end{bmatrix}$

Therefore the standard deviation of the whole portfolio is: $Std(P) = \sqrt{w' Cov \text{ }w} = 0.0866$.

Now what is the correlation of the first asset ($A_1$) with the portfolio?

Well it is given by:

$\rho_{P,A_1} = \frac{Cov(A_1,P)}{std(A) std(P)} = \frac{Cov(A_1,P)}{std(A) std(P)} = \frac{w_1 cov(A_1, A_1) + w_2 cov(A_1, A_2) + w_3 cov(A_1, A_3) +w_4 cov(A_1, A_4)}{std(A) std(P)} = 0.57$

Then just repeat the last step for the other assets (2, 3 and 4).

$\endgroup$
  • $\begingroup$ As you can see from this example, the general rule for the correlation between the portfolio and the i-th security is: take the inner product between the weight vector and the i-th row of the covariance matrix, then divide by the std dev of the i-th security and the std dev of the portfolio. (The former is the sqrt of the [i,i] entry in the covariance matrix, the latter is $\sqrt{w^{'} \Sigma w}$). $\endgroup$ – Alex C May 24 '18 at 11:41

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.