Markman Capital Insight

Popular AI-Powered ETF Suffers From Human Flaw

To learn how to improve your results in the market dramatically by buying options on stocks like Ford and Tesla, take a two-week trial to my special service, Tactical Options: Click here. Members have made more than 5x their money this year.

It is a brand new year and analysts are scrambling to promote their top picks. EquBot, an exchange traded fund manager, is out with its list for 2022 with a twist: the stock picker is not human.

Artificial intelligence is the driving force at its AI Powered Equity ETF (AIEQ). If its algorithms are correct, 2022 will be another great year for the biggest companies with strong momentum.

As with all top pick lists, investors should tread carefully.

EquBot was born in 2017 in the halls of the Haas School of Business at the University of California at Berkeley. Chada Khatua, Art Amador and Chris Nativdad thought they could start a business that would use AI to actively manage a stock portfolio.

As part of the “Startup with IBM,” program the team had access to Watson, the AI that gained fame by beating a steady stream of human contestants on the TV trivia game Jeopardy. Taking Watson to the stock market seemed like a no-brainer. So, the EquBot team got to work training an AI model to automatically parse data from social media, news items, financial statements at the Securities and Exchange Commission, and classic market information such as price momentum and volatility.

The goal was to get the AI to a point where the system could build a dynamic, yet data dependent portfolio that could be rebalanced every day in real time. Ideally, the algorithms would continually learn, getting better as the software systems analyzed more data and accessed its performance.

When the AI Powered Equity ETF debuted in October 2018 it was the first fully functioning, actively managed ETF governed by unsupervised algorithms.

Documents filed at the SEC claimed the AI would analyze more than one million data inputs daily, building dynamic models for its investments. Amador later told CNBC that the system will strive to recognize patterns across data inputs that humans might miss. And given its data dependency, there would be none of the emotional biases that make stock picking so difficult.

Early performance was choppy, to say the least.

The computer system traded as frequently as it did erratically. The turnover rate in the first year was 260% versus only 3.1% for the S&P 500. And during the 2018 market downturn AIEQ plummeted 16%, compared to a loss of 6% for the benchmark S&P. Despite the sketchy start, performance improved and the machine versus man scheme caught on with the public.

The fund logged returns in 2019 and 2020 of 31.2% and 25.4% respectively, earning a three-star rating with Morningstar. Assets under administration grew from only $7 million in 2017, to $169 million through December 2021.

AIEQ enters 2022 with top picks Advanced Micro Devices (AMD), Palo Alto Networks (PANW), DexCom (DXCM), Fortinet (FTNT), Moderna (MRNA), Avantor (AVTR), CBRE Group (CBRE), Enphase Energy (ENPH) and Nutanix (NTNX), according to ETF Managers Group, an ETF tracker and broker dealer.

The list is filled with big cap tech winners in the semiconductors, cybersecurity and vaccine categories. It’s a great strategy if the best trends of 2021 persist. Unfortunately, the list assumes AIEQ will stick around to see how it all plays out in 2022. Given past practices that is not likely.  

A year ago U.S. News and World Report noted that AIEQ to picks or 2021 were Tesla (TSLA), AMD, Enphase, Alphabet (GOOGL), Moderna (MRNA), Zscaler (ZS) and Etsy (ETSY). A year later only three of these companies are in the top nine. And AIEQ gained only 19.4% in 2021, well short of the benchmark S&P 500 advance of 27.6%. 

AIEQ’s algorithms are great at finding winners. The system simply can’t stay with them, at least the way it is currently coded.

Even with all of the data crunching and machine learning, this is a shockingly human shortcoming.

Top-ten picks lists are entertaining, yet most of the time they are not very useful. Investors should find a strategy that works for them, then stay with it.