How a hedge fund robot outsmarted its human master

Japan’s Simplex fund reflects growing use of artificial intelligence in financial trading

When vote counts signalled Brexit was about to become a  reality, a burst of selling sent Japanese shares to their biggest drop in five years. Photograph: Daniel Leal-Olivas/AFP/Getty Images
When vote counts signalled Brexit was about to become a reality, a burst of selling sent Japanese shares to their biggest drop in five years. Photograph: Daniel Leal-Olivas/AFP/Getty Images

Yoshinori Nomura felt like weeping. It was the morning of June 24th, 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 signalled 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 per cent 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 (€31 million) in the fund, one of the first in Japan to utilise artificial intelligence technology.

READ MORE

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 forwhat 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.

Turbulent markets

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 US. The benchmark Topix index has tumbled 16 per cent this year, following an almost 10 per cent rally in 2015. It climbed 0.6 per cent on Monday.

The tumult has been rough on hedge funds, with a gauge of Japan-focused managers tracked by Eurekahedge Pte dropping 3.5 per cent this year. That’s the worst performance since the global financial crisis and compares with a 2.6 per cent 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.

Simple question

Nomura’s strategy appears to blend elements of quantitative analysis and AI, said Kiyoshi Izumi, a professor at Tokyo University who has written a book on the technology’s investment applications.The Simplex fund, like many of its peers, crunches a mind-boggling amount of data to answer a simple question: buy or sell? Nomura’s software focuses on indicators of momentum and trend deviation, making a decision twice a day on whether to purchase or sell futures on the Topix index. It also determines the position’s size, with a cap at 50 per cent of fund assets.

If his program works as designed, Nomura says, its predictive power should improve over time. And while it’s difficult to draw any conclusions after just a few months of trading, the early results are promising.

Nomura’s fund is up 1.9 per cent from its inception in April through August 19th, and he’s targeting annual returns of 7 per cent.

While Nomura's fund is starting small, his firm is no lightweight. Simplex Asset Management is one of Japan's fastest-growing money managers, overseeing 560 billion yen for clients. Despite the challenges of selling an unfamiliar strategy, Nomura expects his fund's assets under management to double by the end of the year amid interest from regional banks and insurers. At least one major investor in Japan is watching the space closely. Japan Post Bank Co's $2 trillion investment unit is interested in AI technology and is aware of Simplex's fund, according to Naohide Une, managing director of the bank's division overseeing hedge fund investments.

"Somebody like Simplex is well positioned," said Bartt Kellermann, the founder of Global Capital Acquisition, which organises "Battle of the Quants" conferences for computer-driven investors in the US, Europe and Asia. "As every day goes by, the machine becomes more intelligent. It's going to be a brave new world."

Bloomberg