Feature request: [Hierarchial Risk Parity Model]

[adityaabothra]

I am making a Hierarchial risk parity Model as my project.
Since it was on your roadmap, will like to create a pull request.
I definitely will take some time and inputs from you, but I believe
I should be able to complete a Weight distribution for assets for diversification.
Ping if needed

[robertmartin8]

Hi @adityaabothra

HRP has been implemented (see docs here), but I'm quite interested in improving it to allow users to specify expected returns.

Robert

[adityaabothra]

Correct if I am wrong in understanding your requirement.
You want user to specify return expected for a corresponding risk.
my query would be
1.Total risk will be given by user, or calculated by the corresponding return entered by user?
2. does this change has to be just a method separately calculated or has something to change within the risk parity clustering, like will such return affect the risk parity on its own?

[robertmartin8]

Really sorry for the slow response.

HRP essentially works as follows:

1. Compute a distance matrix from the covariance matrix
2. Form hierarchical clusters based on the distance matrix
3. Starting from the bottom of the tree, allocate to assets in order to make local minimum-variance portfolios, then proceed upwards (alternatively, can think of it as top down recursive).

The boldened part of step 3 is what I would like to enhance. Rather than just making local minimum-variance portfolios, can we incorporate expected returns, for example, by making local max-sharpe portfolios?

The local allocation is done here, so in principle, we could do something like:

alpha = 1 - first_variance / (first_variance + second_variance)
# change to
alpha = 1 - first_sharpe / (first_sharpe + second_sharpe)

Though we can probably do better by writing down the analytical expression for unconstrained 2 asset max-sharpe.

Speaking of constraints, that would be another major improvement to this HRP implementation. Is it possible to "naturally" put weight constraints on the HRP allocation? Seems quite difficult owing to the recursive nature of allocation, but there are some papers (like this) that describe how it might be done.

Best,
Robert

[adamhwang]

+1 to HRP weight constraints
+1 to local max-sharpe portfolios

[adityaabothra]

Thank You for the response. I will get back to you as soon as possible. Regards Aditya

…

On Mon, 30 Aug 2021, 3:30 pm Adam Hwang, ***@***.***> wrote: +1 to HRP weight constraints +1 to local max-sharpe portfolios — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#200 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHRBYWLFHBM34SSNTIFNAXTT7PL3TANCNFSM4SFHBHOA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.