Lessons from financial experience with artificial intelligence
W.Look Who was the first to adopt new technology? Cutting edge tends to be expensive. Early adopters also tend to be driven by fierce competition to disrupt the status quo. So perhaps no group is more likely to get new tools than the very rich and competitive hedge fund industry.
This rule also seems to apply to artificial intelligence (Love) and machine learning were first adopted by hedge funds decades ago, well before the recent hype. First came the “quants,” or quantitative investors, who use data and algorithms to pick stocks and place short-term bets on which assets will rise or fall. Two Sigma, a New York quantitative fund, has experimented with these techniques since its inception in 2001. Man Group, a British organization with a large quantitative arm, launched its first machine learning fund in 2014. aqr Capital Management of Greenwich, Connecticut Love almost simultaneously. Then came the rest of the industry.Hedge fund experience shows Lovehas shown the ability to revolutionize businesses, but it also shows that it takes time and progress can be disrupted.
eye Funding for machine learning looked like the final stage of the robotic march. Inexpensive index funds, where stocks are selected by algorithms, were already growing in size, with assets under management surpassing those of traditional active funds in 2019. human involvement. Founded in 1982, the flagship fund of Renaissance Technologies, the first ever quantitative investment firm, has earned an average annual return of 66% for decades. In the 2000s, high-speed cables gave rise to high-frequency market makers such as Citadel Securities and Virtu, which allowed stocks to be traded in nanoseconds.new quant outfits like aqr And Two Sigma has beaten human returns and devoured assets.
By the end of 2019, automated algorithms took both sides of the trade. In many cases, high-frequency traders faced off against quant investors who automated the investment process. Algorithms managed the majority of investors’ assets in passive index funds. And all of the largest and most successful hedge funds used quantitative techniques, at least to some degree. Prominent investor Philippe Jabre accused computerized models of “unwittingly replacing” traditional actors when he closed his fund in 2018. As a result of all this automation, the stock market has never been more efficient. Execution was lightning fast and cost very little. Individuals can invest their savings at a fraction of a penny for every dollar.
Machine learning had the potential to deliver even greater results. One investor explained that quant investing began with a hypothesis—the momentum hypothesis—that stocks that have risen faster than the rest of the index will continue to rise. This hypothesis allows individual stocks to be tested against historical data to assess whether their value will continue to rise. In contrast, machine learning allows investors to “start with the data and look for hypotheses.” That is, the algorithm can decide both what to choose and why to choose it.
But the great advances in automation have not continued unabated. Humans fought back. By the end of 2019, all major retail brokers, including Charles Schwab, e*trade and td Ameritrade cut fees to zero in the face of competition from new entrant Robinhood. A few months later, fueled by pandemic boredom and stimulus measures, retail trade began to surge. The day, which peaked in the frenetic early days of 2021 and is calibrating on social media, saw prices skyrocket as his traders piled into unloved stocks. At the same time, many quant strategies seemed to be stalling. In 2020 and his early 2021, most quants underperformed the market and human hedge funds. aqr Closed a handful of funds after sustained outflows.
Many of these trends reversed when the market reversed in 2022. Retail trade share fell as losses piled up. The quants are back with a bang. aqr‘s longest-running fund returned a whopping 44% despite a 20% market decline.
This zigzag, and the growing role of robots, has lessons for other industries. The first is that humans can react to new technology in unexpected ways. Falling trade execution costs seemed to empower the investment machine, but until costs hit zero, the retail industry renaissance was accelerating. Even if the share of private transactions has not peaked, it remains high compared to 2019 and earlier. Private trading now accounts for one-third of his share trading volume (excluding market makers). The advantage of stock options, a type of derivative bet on stocks, is even bigger.
Second, not all technologies make markets more efficient.one of the explanations about of aqr The company’s co-founder, Cliff Asness, argues that periods of poor performance are how extreme valuations have become and how long “every bubble” has lasted. In part, this may be the result of retail investor overenthusiasm. “Having information and having it right away doesn’t mean handling it well,” he thinks Asness. “I tend to think that things like social media make the market less efficient instead of making it more efficient…People don’t listen to dissenting opinions, they listen to themselves. And with politics that can lead to dangerous insanity, and markets that can lead to dangerous insanity, it leads to really bizarre price movements.”
Third, it takes time for the robot to find its place. Machine learning funding has been around for a while now and it looks like it has at least a little edge over its human competitors. However, they have not amassed huge assets, partly because they are hard to sell. After all, few people understand the risks involved. Those who have devoted their careers to machine learning are acutely aware of this. To build trust, says Greg Bond of Man Numeric, the quantitative arm of Man Group, “explain to your clients why you think your machine learning strategy does what they do. We have invested more in
There was a time when everyone thought quants solved it. That is not today’s perception. At least when it comes to the stock market, automation was not the winner-takes-all event many feared elsewhere. It’s like a tug of war between man and machine. Machines have triumphed, but humans have not let go. ■
Sign up for expert analysis of the biggest news in economics, finance and markets. money talka weekly subscriber-only newsletter.
https://www.economist.com/finance-and-economics/2023/03/09/lessons-from-finances-experience-with-artificial-intelligence Lessons from financial experience with artificial intelligence