Calculating Portfolio Volatility and Global Minimum Variance Weights
Summary
This discussion explains how to calculate a portfolio’s expected return and volatility, then outlines how to find the global minimum variance allocation between two stocks. Expected portfolio return is the weighted sum of constituent expected returns. Portfolio volatility requires calculating variance from each asset’s variance and the pairwise covariances, then taking the square root. Covariance is obtained from correlation multiplied by the two assets’ standard deviations; weights enter the variance formula both squared and in cross terms.
For the two-stock minimum variance problem, the response recommends minimizing the portfolio variance expression subject to the allocation weights. It does not provide a completed numerical answer or the closed-form solution, and the referenced equal-volatility setup is a hint rather than a full derivation. The method assumes the supplied volatility figures are standard deviations and depends on having the correlation between the two assets. Simply weighting individual volatilities does not generally give portfolio volatility, and return-based weights do not identify the minimum variance portfolio.
Key ideas
- Expected portfolio return is the sum of each asset’s expected return multiplied by its portfolio weight.
- Portfolio variance includes weighted asset variances and covariance terms for each asset pair.
- Covariance equals correlation multiplied by the two assets’ standard deviations.
- Portfolio volatility is the square root of portfolio variance.
- Minimum variance weights are found by minimizing portfolio variance subject to the allocation constraint.
Tags
Full text
# Portfolio returns, volatility and weights of capital
# Portfolio returns, volatility and weights of capital
I would just like to check if I've done these questions right, I feel like I might have used the complete wrong methods to get my answers.
I've been given information on 3 stocks:
I've filled in the missing values in the correlation matrix and I've been asked the following questions:
- "An asset manager forms a portfolio by investing 10% of capital in Stock A, 40% of capital in Stock B, and the rest in Stock C. Calculate the expected return and volatility of this portfolio."
- " If you are a portfolio manager, your client requests to invest in the global minimum variance portfolio using Stock B and Stock C. Calculate the weights of capital for this portfolio."
So for Part 1 this is what I did:
I took the expected returns and multiplied them by their percentage in the portfolio; so I had $A = 8 * 0.1 = 0.8$% , $B = 10 * 0.4 = 4$%, $C = 15*0.5=7.5$% and the total expected return of the portfolio would be the sum of these values, which is $12.3$%
For the volatility I used the same method, multiplying the volatility in the table by the percentage of the stock in the portfolio and I got the total volatility as $14.5$%
I'm not sure if my method is correct so please let me know if what I've done is ok or completely invalid.
For Part 2:
I wasn't entirely sure what to do but I added the expected returns of $B$ and $C$ and got $25$%, $100/25 = 4$ so I thought the weights of each stock would be $Weight B= 10*4 = 40$% and $Weight C=15*4=60$% but I'm not sure if this is correct, should I have used the volatility instead since the global minimum variance portfolio cares more about volatility than return so the weights would be $50/50$ ? Or is there a completely different method I should use?
These questions might be very simple for this site but I would just really like to know if my methods are ok or if I'm just making up methods that aren't valid, any help would be appreciated.
## Answer by mark leeds (score 0, accepted)
https://quant.stackexchange.com/a/64120
Hi: I'm not sure how the question defines volatility but let's assume it's standard devation so $\sigma_{i}$ for the ith asset.
So, if you want the volatility of a portfolio with weights, $w_1$,$w_2$ and $w_3$, you need to calculate the portfolio variance and then take the square root.
To get the portfolio variance, it's
$var(w_1 X_1 + w_2 X_2 + w_3 X_3) = $
= $w_1^2 var(X_1) + w_2^2 var(X_2) + w_3^2 var(X_3) + 2 w_1 w_2 cov(X_1,X_2) + 2 w_2 w_3 cov(X_2,X_3) + 2 w_1 w_3 cov(X_1,X_3)$
( check the formula because it's been a while and my memory fails me sometimes ).
Note that, in the expression above,
A) the var terms are the squares of the volatilities given.
B) the cov terms are the respective correlations in the matrix multiplied by the the given volatilities.
So, cov( $X_1,X_2) = \rho_{1,2} \sigma_1 \sigma_2$.
C) The $w_{i}$ are the weights so 0.10, 0.40 and 0.50
Then, when you put those in, take the square root of the result and that's the volatility of the portfolio.
For part 2), here's a hint. If you had a portfolio of stocks with equal volatility of 15 percent, then the variance would be:
$var(w_2 X_2 + w_3 X_3) = w_2^2 Var(X_2) + w_3^2 Var(X_3) + 2 w_2 w_3 cov(X_2 X_3) $.
So, part 2 is asking you to find the $w_2$ and $w_3$ that minimize the expression above, given that the var terms are 0.15^2 and cov term is known ( for cov take correlation of B and C and multiple by their volatilities ). Oddly enough, I was once asked the same question on an interview so I found this question interesting.Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.