4
$\begingroup$

If I have three asset classes and their historical weekly returns for five years, how can I construct a minimum variance portfolio and an efficient frontier plot with Excel? To do that do I have to assume the return is normally distributed?

Update: there's a host of tutorials to plot the frontier for two assets as long as I have a table of say 10 possible weights of one asset. But with three assets or more plotting would be challenging as I have to come up with a much bigger table for the combination of the weights of these assets. As such I was wondering if there is some kind of simulation algorithm or any techniques to make it easier.

$\endgroup$
  • $\begingroup$ lmgtfy.com/?q=minimum+variance+portfolio+excel $\endgroup$ – vonjd Jan 4 '15 at 21:14
  • 1
    $\begingroup$ Hi nouveau, welcome to Quant.SE, as vonjd points out you have to show more effort to be on-topic here. $\endgroup$ – Bob Jansen Jan 4 '15 at 22:01
  • 1
    $\begingroup$ Hi Bob, sorry that I wasn't specific enough. As vonjd points out, there's a host of tutorials to plot the frontier for 2 assets as long as I have a table of say 10 possible weights of one asset. But with three assets or more plotting would be challenging as I have to come up with a much bigger table for the combination of the weights of these assets. As such I was wondering if there is some kind of simulation algorithm or any techniques to make it easier. $\endgroup$ – nouveau Jan 4 '15 at 22:18
  • 1
    $\begingroup$ @nouveau: Then please edit your question accordingly and it will be considered for reopening :-) $\endgroup$ – vonjd Jan 5 '15 at 8:49
1
$\begingroup$

Yes, the assumption in MPT is normal distribution for returns.

You can programme yourself in R or Excel, following elementary linear algebra.

Eric Zivot (U Wash) has a spreadsheet solution here:

$\endgroup$

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.