Could fund strategies be better managed by Intelligent ML/Ai based platforms?


#1

I would be interested in knowing what your views are around a platform which could run any fund strategy and optimise it for Portfolio performance metrics? If such a platform were to exist, it would offer Asset Managers the ability to focus on AUM whilst the the Machine Learning based automation takes care of the fund management and produces optimal performance. This would fundamentally alter the value offering - humans are great at relationships ie managing inflows of AUM, machines are better at understanding data and automation - ie processing market price data to evaluate optimal investing strategies and returns.


#2

Hi @tomn welcome to Fintech Genome. Is what you have in mind something at the reporting layer, for example to track style drift of managers?


#3

Hi @BernardLunn thanks for welcoming me, it is a great forum you have here. I wasn’t necessarily referring to style drift of managers, but yes a if a platform were to run a strategy it would be more consistent with the defined methodology and be self directed. However, what I was alluding to was, if an asset management platform existed which could manage Fund strategies and allow the manager to simply optimise the strategies for performance or volatility management or some other improvements. Such a platform could run multiple strategies and be scalable, something which is currently limited by human capacity.

How would the world see it?


#4

@tomn at first glance it sounds like quant trading, something like Renaissance. Is that right? What about Quantopian or Numerai? Ask @Efi she knows this space very well.


#5

@tomn the two examples that @BernardLunn mentioned qualify. They are different approaches that adress different biases that are inherent in human management of assets.

Check Quantopian

and Numerai


#6

@Efi these are brilliant examples of where the strategies are being democratised, and I suppose they work on a model which is based around marketplace sourcing of algorithms and harvesting the best algorithms which may apply to a certain aspect of the market. This is a clearly a fascinating area and kudos to the ones on this crowdsourcing journey as it certainly has purpose.

The only challenge if any, I see is how do they address each Asset Manager’s individual strategy around Portfolio construct methodology, rebalancing rules,volatility management, drawdowns and other ratios.

This leads me on to the point what @BernardLunn was making that it sounds like what Renaissance are running, a fully managed and configurable platform for running any Portfolio strategy. Yes @BernardLunnyou are right, that is what I was alluding to. This is the view, we subscribe to, where a platform has enough intelligence to provide the performance but also the capability for further refined optimisation. The base layer is strong and tactile enough to allow porting of any Portfolio strategy yet be flexible for optimisation if such needs arise.

This would then change the landscape, give more autonomy to Asset Managers to manage AUM optimally whilst the platform manages the strategies which are individual and integral to those Asset Managers. To some extent that is one way, the future of the industry may move too.

I want to keep an open mind, and somehow I still feel that in the new world of ML/AI based Asset Management, both the data scientists, the machines and importantly smarter Asset Managers have to co-exist. The question is around a platform which will optimise these relationships.