Let say I have two strategies, A and B. Historically, when the portfolio value of strategy A moves up, then that of portfolio B moves up. Same in the down case. Then, we can say both strategies are highly correlated each other (close to 1).

What I want to do is to choose some uncorrelated strategies among the strategies pool so that overall(maybe equal-weighted stratgies) portfolio's variance can be minimized.

However, if the magnitude of movement in one strategy is greater than the other, then they can till drift apart event though they are correlated ...

Here is a simple example:

enter image description here

Both cases show the same correlation values, but if I manage both strategies with equal weights, the final portfolio values will be totally diffrent.

So, I decided not to see the correlation of return when selecting strategies. Is there any substitution?

  • $\begingroup$ correlation only measures linear co-dependence though. There could also be non-linear co-dependence between strategies. Besides this, what you're looking for are called ensemble models $\endgroup$ – develarist Dec 2 '20 at 16:56

Your Answer

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

Browse other questions tagged or ask your own question.