Machine learning for actively-managed ETFs

It’s been one year since my landscape post on AI in Digital Wealth mgt: Sniffing out investment opportunities and the conversation continues on the Fintech Genome with the addition of new ventures and with other insightful conversations around whether Could fund strategies be better managed by Intelligent ML/Ai based platforms?

This is all happening while the rise of passive investing continues and the push for non-active ETFs continues. Statistics show that human stock picking is in decline, with Blackrock showing 64% of its assets under management in passive-focused index and exchange-traded funds and only 29% in funds that follow active strategies.

AI in digital wealth management clearly isn’t about passive and indexing. It is about discovering new sources of alpha, managing risk in a better way or about democratizing access to certain kinds of alpha. AI in wealth management is always about extracting value from data, whether it is natural language data, sentiment analysis, or numerical data processing.

Today I’ll zoom into a partnership between two AI&ML analytics providers and two Fintechs, that aims to capture alpha and offer it to all of us, in the most popular wrapper, the ETF.

Watson is not new to financial services. It is behind various IBM initiatives (e.g. IBM Cloud for financial services) and it is powering already some Fintechs, like Alpha Modus, focused on asset management services using cognitive analytics. None of these ventures have packaged financial instruments ready to be added to our portfolio.

The AIEQ ETF to be listed on NYSE

The AIEQ ETF pending its approval will be listed on the NYSE. It is the brainchild of a triangular collaboration between Watson + TensorFlow, Equbot, and the ETF Managers Group.

Watson shares its ability to parse text (news, articles etc) to uncover catalysts that can affect (positively or negatively) the individual stocks and real estate trusts under consideration.

Equbot, is the US based Fintech that sprung out of the IBM Global Entrepreneur roaster. It leverages the Watson AI abilities, to build predictive models that combine fundamental analysis and Machine Learning. It uses the open source Machine Learning libray TensorFlow to build the Equbot model. This model identifies 30-70 US stocks on a daily basis that have the greatest appreciation potential.

The ETF Managers Group (ETFMG), is the largest private ETF issuer/manager in the US, and is managing the process of filling for an ETF to be listed on the NYSE. The ETF is named the Equbot with Watson AI Total US ETF (AIEQ). Equbot is the subadvisor to this fund. Through its partnership with Watson and TensorFlow, it will develop the Equbot model which is driving the actively managed fund decisions.

Actual risk-adjusted performance can only tell whether stock picking based on the combination of ML software and cognitive analytics, is generating alpha after the taking into account the costs of active trading and managing the ETF wrapper.

Efi Pylarinou is a Fintech thought-leader, consultant, and investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.




The crazy brilliant IBM Adept fork of Ethereum

Another update as of June 2016 in italics

Since writing this an aeon ago (in Sept 2014), much has happened and much has not happened. So this is an update as of end Jan 2015.


– many more people are connecting the dots between re-decentralization and Internet Of Things.

– many more announcements from IBM and Samsung and Ethereum.

Not Happened:

– any shipping products.

The last point will bring out the skeptics, but I am sticking with two convictions:

1. Crazy brilliant move by IBM. The need for decentralization in IOT is immediate but more long term in finance. IBM gets to make money now in IOT and learn more to make money in finance later.

2. Ethereum is the platform to bet on. There is a lot of debate on this score, but the two big reasons why people are skeptical of Ethereum look less compelling with each passing day:

A. Ethereum is not shipping yet. Yes, but it is getting closer every day. Sure, they could still fail to ship working code, but that does look more unlikely with every passing day.It is shipping now. Obviously it is still a work in progress (but so is every tech product that does not die). Today the preponderance of evidence is that Ethereum will work.

B. Ethereum does not use Bitcoin. It uses Ether. Link A + B and you get cries of Scam! Now that Bitcoin price is falling below its mining costs, the idea of a Blockchain platform that divorces Blockchain from Bitcoin looks increasingly smart. (No, I have not bought Ether, so this is not a pump and dump post).Ethereum DAPPS (some examples here) do also use Bitcoin as the crypto currency. It is wrong to look at it as Ether vs Bitcoin – it is more like “horses for courses”.

Original Post in Sept 2014:

This will be a short post, as I am still getting my head around the implications of this announcement that IBM is creating a fork of Ethereum called Adept (news release as pf Jan 2015  here).

I have been fascinated by Ethereum for some time. (Index to all Ethereum related posts on Daily Fintech are here).

I had a thought about how Ethereum connects with Internet of Things in relation to the professionalization of sharing economy services such as AirBnB here. This is worth reading and the need I was expressing there has been realised in (which represents the best use case of decentralised smart contracts in the wild).

All I know now about the IBM announcement is:

  • This is a massive validation for Ethereum. A company of the scale of IBM ($98 billion of revenue last year) does not get involved with bleeding edge technology like this on a whim.
  • This demonstrates what an amazing company IBM is. To be able to operate simultaneously at the level Fortune 500 Board/CXO and at the level of bleeding edge technology that is usually only understood by a few developers is very, very cool.