Artificial intelligence can now pick stocks and build portfolios. Are human managers about to be replaced?

Outside of their ability to understand a company’s fundamentals, one of the skills Raj Lala appreciates most about his portfolio managers is their ability to interpret body language.

Sitting across from management teams before making a decision to either invest or divest from their companies, Lala, the CEO of Evolve ETFs, said his portfolio managers can learn a lot from simply reading the room. Maybe they spot a nervous twitch after a question on guidance or a CEO unable to make eye contact when responding to a question about declining revenues.

That very human capability was at the forefront of Lala’s mind when he was recently pitched on two types of artificial intelligence that he could incorporate into his portfolio management processes. And it’s one of the reasons he said no.

“I can’t see AI getting to that point where it replaces human interaction and, quite honestly, I would say god bless our world if that’s the case,” Lala said.

Despite Lala’s skepticism, AI technology that can manage portfolios already exists to some extent and is beginning to be deployed.

In 2017, San Francisco-based firm Equbot LLC launched its AI Powered Equity ETF, which assesses more than 6,000 U.S.-traded stocks per day and decides where to invest based on its analysis of regulatory filings, earnings, valuations and other fundamentals.

Weeks later, Horizons ETFs launched the first AI ETF in Canada, where the AI is responsible for building a portfolio from 32 global equity ETFs based on its analysis of money flow, volatility and moving averages. Here, however, a human is required to pull the trigger on any trade.

Both ETFs are currently trading below the value at which they first opened in 2017, something that is certain to give fuel to skeptics who believe that while AI has made incredible strides in the last few years, it simply is not yet ready to make the job of a portfolio manager defunct.

One reason, according to Stuart Sherman, is that there are too many variables within stock picking that cannot be programmed into an AI process.

Sherman, the CEO of Toronto-based AI firm IMC Business Architecture, compares the challenge to a report in the Guardian which said that a cat received better investment returns than three portfolio managers and high school students in a 2012 challenge.

In that same time, AI could also outperform humans in the role of a portfolio manager, but it wouldn’t really prove that it was better unless it was able to consistently beat them by double digits over a period of multiple years.

“It could work for a while,” Sherman said. “But the cat will work for a while. Eventually, it’ll regress to the mean.”

Sherman described AI as a pattern recognition tool and said that to build one that could manage a portfolio, a developer would have to start with a “ground truth.” Essentially, a programmer has to show an AI hundreds, or even thousands of examples of what a good portfolio is so that it can trace the pattern. The problem is that an AI’s knowledge is based on past data and it cannot account for the randomness that sends some stocks soaring and others into the dirt.

Take the example of Elon Musk appearing on a podcast and smoking cannabis, he said. The appearance and widely circulated memes of Musk taking a toke led to a six per cent drop for Tesla Inc. An AI system may have picked up on Musk’s behaviour prior to the podcast appearance, such as his feud with the SEC, and make a sell suggestion but “on the other hand, Musk acted radically when he started Tesla,” Sherman said.

Instead of focusing on building portfolios, the Toronto-based IMC Business Architecture is working on AI that would help portfolio managers better select their clients. Before accepting a client, the behavioural AI Sherman and his team have developed would be able to collect language samples from people and put them into clusters of investors who are like-minded.

The AI could then assign those groups of people based on their investment risk, their social goals and personality to the portfolio manager that best suits them. Even a subtle improvement with taking on the right clients could lead to substantial profits, he said.

Like IMC, leading Canadian firm Element AI sees the machines and humans working together and has incorporated that belief into its developmental process, according to chief science officer and co-founder Nicolas Chapados.

Chapados’ team has been focused on perfecting what they call the Trade Flow Scheduler, which is designed to help institutional investors such as pension funds rebalance their portfolios. In order to do so, these investors may be forced to execute larger orders in a market without the needed liquidity. Chapados said Trade Flow Scheduler can analyze market conditions and make recommendations on how many days or weeks the trade should be made in as well as making suggestions on inflows and outflows in that period that would have the least impact on the market.

Asked if Element would one day go further and attempt to build an AI that could replace a portfolio manager, Chapados said he wouldn’t comment on future projects.

“Our goal is to not fully replace human beings but to provide a second opinion, if you will, and to augment the human in the role,” Chapados said.


Filed Under: Technology

About the Author:

RSSComments (0)

Trackback URL

Comments are closed.

Read previous post:
Why B2B brands need to invest in brand marketing

Creativity and storytelling were previously seen as a luxury afforded to B2C brands but B2B marketers are waking up to...