So far, this promise has not been kept. The only major fund managed by AI – AI Powered Equity ETF – underperformed and received little investor interest. Since opening in October 2017, the fund has returned just under half of the Vanguard S&P 500 ETF, 35.3% vs. 70.8%, with slightly more volatility: 24.2% vs. 21.5% . The AI ETF – ticker AIEQ – has a beta of 1 against the S&P 500 after fees, with a negative alpha of 3.82%, meaning it lost 3.82% against the S&P 500 each year on average over the past five years. However, this alpha is not statistically significant, which means that it is plausible that AIEQ had a higher long-term expected risk-adjusted return, but only had an unlucky five-year one.
Another leading artificial intelligence product is HSBC Holdings Plc’s AI Powered US Equity Index, or AIPEX. Since its inception in August 2019, the index has returned just 2.3%, compared to 44.8% for the Vanguard 500 index. However, AIPEX has an annualized volatility target of 6%, or about a quarter of the 24.3% of the S&P 500 over the same period. AIPEX met its volatility target almost exactly, 6.1%, and it has a beta of 0.19 against the S&P 500 and a negative alpha of 1.8% (and like AIEQ, this negative alpha does not is not statistically significant). AIPEX includes an index fee of 50 basis points and holds most of its hypothetical capital in cash. Taking these two elements into account, AIPEX’s pure stock selection – the measure of its AI’s success – has lost 6.8% annually against the S&P 500 for the past three plus years.
Nevertheless, AI is making strong inroads in investment management. The main area is the processing of “unstructured data” such as news stories and textual reports. There is no doubt that AI trumps humans in this regard; he can read everything, in all languages, and distill useful information. It can process images and anything that can be converted into bytes in a computer file. The amount of this data is rapidly increasing and the sophistication of the algorithms to process it, so AI will continue to advance in this task.
Another area where AI and ML have been used extensively is in trading algorithms – not deciding what to buy and sell, but choosing how to split orders and enter them into a variety of trading platforms. These algorithms don’t need to be very smart, their main advantage over humans is speed. They can continuously monitor hundreds of pricing data feeds and make instant decisions.
But these auxiliary functions were not what AI pioneers dreamed of. They believed that AI could take over the entire investment decision process, and not just create signals and execute trades, but also interpret those signals and choose trades to execute. Bryan Kelly, head of AI research at AQR Capital Management LLC (where I used to work), explains it this way:
“Machine learning has a real impact on systematic investment processes because it allows managers to more quickly and expressively metabolize information from more new sources (thanks to greater model flexibility). But it’s important to remember that the central motivation of machine learning — extracting as much usable insight as possible from data — has long been the modus operandi of quantitative investing. So I see ML as another step in the evolution of quantitative investing methods. ”
I think that represents the prevailing belief right now. Artificial intelligence is slowly being incorporated into quantitative investing, especially for signal mining and trading, but it is enhancing human research and decision-making instead of replacing them.
There are two areas of hope for a greater role for ML in investment management. The first is the “L” in ML. Every day of underperformance is another opportunity to improve. Perhaps ML is like a baby bird that has just found its wings and will one day soar high above Earthbound humans. The second is that institutional investors are interested in using ML for asset allocation rather than stock picking. Cross-market optimization is much more difficult than selecting portfolios within asset classes. Most investors don’t even dabble in it, rather they build the best portfolios of stocks, bonds, and commodities they can, and so on, then combine them based on pre-selected allocations. AI is the only known approach to building a true global portfolio.
Investors should forget about looking for a Skynet or a Hal 9000 to manage their money for now. Top companies are using ML where it has been proven to work – and perhaps considering other applications – but pure ML decision making has lagged in the market.
More from Bloomberg Opinion:
• Creative AI generates messy problems: Parmy Olson
• AI can help make cryptocurrency safer for everyone: Tyler Cowen
• Our future AI overlords need a resistance movement: Parmy Olson
This column does not necessarily reflect the opinion of the Editorial Board or of Bloomberg LP and its owners.
Aaron Brown is a former Managing Director and Head of Capital Markets Research at AQR Capital Management. He is the author of “The Poker Face of Wall Street”. He may have an interest in the areas he writes about.
More stories like this are available at bloomberg.com/opinion
#Analysis #beat #market #days