As the robot war on Wall Street stock pickers heats up, there's a new line of attack from the algorithmic set: IBM's Watson supercomputer has been hired to help run an ETF and pick stocks than can achieve better performance than the broad U.S. stock market index.
The ETF, called the Equbot with Watson AI Total US ETF, has been filed for by ETF Managers Group, which works with a number of ETF subadvisers to bring new investing ideas into the market, and already has launched big data, cybersecurity, drone and immunotherapy funds, among others.
The Watson ETF's approach to picking stocks is described in the filing with the Securities and Exchange Commission as "actively managed" and "based on the results of a proprietary, quantitative model (the "Equbot Model") developed by Equbot LLC ("Equbot") with Watson."
Equbot, the Fund's sub-advisor, is a technology-based company focused on applying artificial intelligence to investment analyses. It is part of the IBM Global Entrepreneurs start-up roster. IBM already has a Watson effort for financial services more broadly, which includes a Watson analytical tool for wealth advisors and wealth management groups, and Watson applications for financial markets analysis.
The filing says Equbot will use IBM's Watson AI to perform a fundamental analysis of U.S.-listed stocks and real estate investment trusts based on up to 10 years of historical data and then apply that analysis to recent economic and news data.
"Each day, the Equbot Model ranks each company based on the probability of the company benefiting from current economic conditions, trends and world events and identifies approximately 30 to 70 companies with the greatest potential for appreciation and their corresponding weights, while maintaining volatility comparable to the broader U.S. equity market."
One of the most successful examples of algorithmic stock-picking in the history of Wall Street is hedge fund titan Robert Mercer, co-CEO of Renaissance Technologies, one of the most profitable hedge funds in the world. Mercer came to Renaissance in 1993 from IBM, where the computer engineer did pioneering work on using computers to review massive amounts of text and then use predictive analytics to translate between languages, an algorithm that laid the groundwork for Google Translate and Apple's Siri.
"The trend with ETF product development is toward quantitative efforts following predetermined rules to ensure consistency. It seems logical that more efforts will involve computer programming going forward using back-tested tools," said Todd Rosenbluth, director of mutual fund and ETF research at CFRA.
Neena Mishra, director of ETF research at Zacks Investment Research, said she likes the idea, but while AI can be used to process and analyze vast amount of data much quicker than humans, sometimes the challenge lies in deciding the importance of each piece of information in the investment decision. "An investment process involving a human analyst/team of analysts, supported by strong data analytics, certainly makes sense," she said.
She also noted that the ETF's expense ratio has not been disclosed, and since it's actively managed, it could be high. "That's the main reason why I don't like most actively managed funds. Active managers' performance has been underwhelming in general and does not justify high management fees charged by them."
BlackRock, the world's largest money manager, with more than $5 trillion in assets — and owner of the iShares family of ETFs — recently decided to turn over management on many of its actively managed funds to algorithms.
In many markets, BlackRock's automated trading products have beaten indexes more consistently than human fund managers, but they suffered a hiccup in 2016. BlackRock CEO Larry Fink told CNBC in April that the professionals who had been expected to be cut would be shifted to other jobs that make more use of analytics.
In the past, BlackRock has explained that there are some tasks only a computer can do when it comes to analyzing large sets of data to make stock calls. That includes monitoring satellite data of big-box store parking lots and analyzing internet searches for consumer products to predict sales volume or even national economic growth.