Re:Modern Portfolio Theory: A Case Study on Turnips
by hueyduck on Monday June 01, 2020 @10:46PM
(#23386792)
Thanks for sharing my post here. There's actually one more thing I tried but didn't make it into the post - if we treat the 12 trading windows as distinct asset classes characterized by their own expected return/volatility plus the pair-wise correlations, it's possible to construct tangent portfolio/efficient frontier and get the optimal weights. The negative expected return on a few of them certainly makes the exercise less meaningful though.
[ Parent
| on HN ]
Re:Modern Portfolio Theory: A Case Study on Turnips
by shocks on Tuesday June 02, 2020 @06:19AM
(#23389484)
> There's actually one more thing I tried but didn't make it into the post - if we treat the 12 trading windows as distinct asset classes characterized by their own expected return/volatility plus the pair-wise correlations, it's possible to construct tangent portfolio/efficient frontier and get the optimal weights.
This was recently explored here [1] and here [2]
Re:Modern Portfolio Theory: A Case Study on Turnips
by hueyduck on Tuesday June 02, 2020 @09:02AM
(#23390519)
Excellent point - I've actually read [1] but not [2]. It's worth noting that neither of them suffers from the same negative weights problem that I am seeing, which is probably due to the fact that 1) an explicit bound between 0 and 1 is imposed on the weights 2) (to a much lesser extent) the analysis is conducted conditional on a specific realized path as of Wednesday.