Yoshinori Nomura felt like weeping. It was the morning of June 24, Brexit day, and markets were moving against him.
Well, not against him, exactly. It was the hedge fund manager’s self-learning computer program that had placed the bet, selling Japanese stock-index futures before a sizable market advance. Nomura had anticipated a rally, but decided not to interfere, and his fund was paying the price.
Then, in an instant, everything changed. When new vote counts signaled Britain was going to leave the European Union, a burst of selling sent Japanese shares to their biggest drop in five years. By luck or design, Nomura’s Simplex Equity Futures Strategy Fund ended the day with a 3.4 percent gain, one of its best results in three months of trading.
“The machine was right after all,’’ said Nomura, who spent more than three years refining his trading program and now oversees about 3.5 billion yen ($35 million) in the fund, one of the first in Japan to utilize artificial intelligence technology.
Nomura doesn’t have the assets or name recognition of computer-savvy giants like Renaissance Technologies or Two Sigma Investments. But in his own way, the Tokyo-based physics buff has become a compelling test case for what some say is the future of money management. If Nomura can succeed in Japan -- where central bank stimulus has upended markets, hedge funds are trailing global peers and institutional investors are notoriously risk averse -- it would offer hope for fledgling AI traders around the world.
The 43-year-old money manager is setting his software loose in one of the planet’s most turbulent markets. Share-price swings in Japan rank No. 1 among the world’s 15 largest stock venues, with volatility readings almost four times higher than in the U.S. The benchmark Topix index has tumbled 16 percent this year, following an almost 10 percent rally in 2015. It climbed 0.6 percent on Monday.
The tumult has been rough on hedge funds, with a gauge of Japan-focused managers tracked by Eurekahedge Pte dropping 3.5 percent this year. That’s the worst performance since the global financial crisis and compares with a 2.6 percent of gain for the research firm’s index of managers worldwide.
It makes for a difficult backdrop as Nomura tries to drum up appetite for a strategy with almost no real-world track record. Some prospective institutional clients are so hesitant to stick their neck out for an unfamiliar product that they asked Nomura to remove the term AI from Simplex’s promotional materials. They didn’t want to have to explain how it works to their bosses, he said.
“There are very few AI funds currently out there globally,” said Mohammad Hassan, a senior analyst at Eurekahedge in Singapore. “These early adopters of machine-learning methodologies will need to build a good track record before investors give them serious attention.”
It’s often difficult to distinguish AI funds from their more ubiquitous "quantitative" investing precursors, according to Motoyuki Sato, a general manager and researcher at Man Group Japan Ltd., a unit of the world’s largest publicly-traded hedge fund manager. While they both rely on computers to make investment decisions, AI programs go a step further than quant software by attempting to improve themselves over time -- mimicking the human brain’s capacity for learning.