Would you let a thinking robot pick stocks for you?
Fund companies are increasingly turning to artificial intelligence to help inform their investment-management decisions. Using AI to manage investments is nothing new for institutional investors—some of the world’s most high-profile hedge funds including Renaissance Technologies and Bridgewater Associates rely on computers’ analysis to make day-to-day trading decisions. But AI-powered investing is making its way to individual investors, as well.
In October, two new high-profile products using artificial intelligence were launched— AI Powered Equity ETF (AIEQ) and the Credit Suisse RavenPack Artificial Intelligence Sentiment (AIS) Index.
AIEQ is the first ETF powered by International Business Machines Corp.’sIBM -1.15% Watson cognitive computing platform and is managed by San Francisco-based EquBot, an investment company with plans to launch several AI-powered ETFs in the future.
Credit Suisse AG joined with New York-based RavenPack on its AI-powered index product. The investment bank says artificial intelligence will be used to make sector-weighting decisions in the index, which tracks the performance of U.S. large-cap stocks. The AIS index is the first AI product for Credit Suisse.
A machine’s insight
Both AIEQ and AIS work on the same principle—they rely on machines to analyze reams of analyst notes, articles, earnings-call transcripts, regulatory filings and social-media posts on individual companies to gather insights about how they might perform. From there, the software determines the weighting of their respective indexes and trades.
Art Amador, co-founder and chief operating officer at EquBot, says that EquBot’s algorithms with the power of Watson will look for undervalued stocks to invest in that have the potential to outperform over time. A human will monitor the decisions to ensure data quality.
AIEQ represents an evolution in the investment process and landscape. Mr. Amador says his company chose to use Watson because the platform works well with investments.
“All AI is not created equal,” he says. “We wanted to be able to offer transparency into what stocks were being chosen and why they were being chosen. The EquBot approach to using AI for investing is designed to provide that understanding.”
Watson is already being used in financial services and banking for regulatory compliance and financial data analysis. AIEQ is the first ETF powered by Watson.
Mutual-fund managers are also bringing artificial intelligence into the investment process.
In October, AXA Investment Management’s Rosenberg Equities announced it was taking steps to integrate AI into its investment products. Fund managers at Rosenberg Equities, which already uses algorithms to run its funds, also plan to use natural-language-processing technology to analyze things like analyst notes, articles and other unstructured data that could have an impact on how stocks perform. Rosenberg says natural-language processing will inform its smart beta and “core plus” strategies, which are designed to beat the performance of benchmark indexes, largely through stock weightings.
According to Michael Kollo, head of research and chief quantitative strategist for Rosenberg Equities at AXA IM, artificial intelligence will also be folded into investment modeling to help fund algorithms predict risk and extreme price movements in specific stocks.
“We already have a significant computing infrastructure in place, so it makes sense for us to take this step and continue to improve our process,” Dr. Kollo says.
The bond bots
AI is also creeping into fixed-income funds. New York-based Wavelength Capital Management’s Wavelength Interest Rate Neutral Fund (WAVLX) has been using a proprietary artificial-intelligence system to analyze information and make trading decisions about the bond market since 2013. WAVLX is designed to capture positive return from the bond market without being sensitive to interest rates.
“When we started, there wasn’t a lot of data available,” says Mark Landis, co-founder at Wavelength. “We were training our software using articles, historical patterns, social-media posts—a lot of unstructured data. As regulation increased reporting requirements for all asset classes, we’ve been able to add that information into our analytics platform. The insights are getting better each year.”
WAVLX is up 4.56% year-to-date through Oct. 27. The Bloomberg Barclays U.S. Aggregate Bond Index, the benchmark index for the bond market, was up 2.90% for the same period.
Is it all over for humans?
Not quite. With all of these products, even though computers are taking on the role of a hyper-dialed-in portfolio manager, there are still humans standing by, ready to flip the switch if the software gets carried away.
Dr. Kollo, the head of research at Rosenberg Equities, cautions that investors should use that same kind of vigilance when choosing AI-powered funds.
“The technology itself is only a tool, like a shovel or a hammer,” Dr. Kollo says. “Success still depends on the skill of the manager running the strategy,” he says. “Investors still need to understand the team, what the process is and what the goals are before they invest.”