SHANGHAI/HONG KONG (Reuters) – China’s algorithm-driven quant funds boomed in 2021 as investors sought alternatives to a languid stock market, but the final months of the year saw some “flash boys” bogged down by heavy volatility and their sheer size.
High-flyer Quant, a top hedge fund house in China that uses powerful computers and artificial intelligence (AI) to exploit market opportunities, last week apologised to investors for a record slump in performance. The fund house, which manages roughly 100 billion yuan ($15.69 billion), blamed the crash on wild shifts in investor sentiment and crowded trades in a sector that is “growing too fast in size.” High-flyer’s public apology echoes an earlier setback in 2021 at Shanghai Minghong Investment Management Co, another quant heavyweight fund that suffered losses following a surge in assets under management (AUM).
“When you’re small, high-frequency, rapid-trading strategies could create very beautiful performance charts,” said Wang Li, chairman of Hangzhou Liberty Fund Management Co. But after big money gushes in, “the strategies no longer work.”
While high-frequency traders, known colloquially as “flash boys”, typically ride on market momentum and gyrations, their trades can become ineffective if there is too much money betting the same way.
When such funds grow too large, they lose their ability to move nimbly and exploit market inefficiencies.
A High-flyer Quant fund, designed to enhance returns on the CSI500 Index, lost 13.1% during the last quarter, despite a 3.6% gain in the underlying index, according to fund distributor Simuwang.com. That wiped out most of the 21.9% gains from the first nine months. Another High-flyer fund using hedging strategies lost 12.1% during the final three months of 2021, bringing the year’s performance to an 8.1% loss. That fate could befall other players in a sector that, according to Citic Securities estimates, has jumped ten-fold over the past four years to 1 trillion yuan. Daily turnover in China’s A-shares market crossed a trillion yuan on most days in 2021. Around 25 hedge fund managers that use such quantitative strategies each saw their AUMs surpass 10 billion yuan last year, according to Kaiyuan Securities. Several, including High-flyer, exceeded 100 billion yuan in AUM. Investors rushed into quant funds, which seek to generate absolute, or enhanced returns, during a year in which regulatory crackdowns on sectors including tech and property roiled stock markets. China’s blue-chip CSI300 Index lost 5.2% in 2021, despite record annual turnover in local stocks, fuelled in part by the rise of rapidly trading quant funds. The problem with top quant players like High-flyer Quant is that “their size is too big,” said Shi Ke, partner at hedge fund manager iFund Asset Management Co. “Another issue is that AI can do a good job analysing historical data, but may not react quickly when market swerves.”
Last year, some clients made returns lower than the benchmark, while some others suffered paper losses, and “we feel very sorry,” High-flyer Quant said in the letter, vowing to invest more in research. SHQX Asset Management, a Shanghai-based hedge fund manager, said many so-called CTA funds, which trade commodities using quantitative strategies, suffered towards the end of 2021 due to slump in commodity prices. There are signs of growing caution in the industry. Several fund managers, including High-flyer, Tianyan Capital and Evolution Asset Management, have stopped taking new money for some or all products.
Breakneck growth has increased regulatory scrutiny.
Top securities regulator Yi Huiman said in September growth in “quants” was a challenge to stock exchanges. Meanwhile, the Securities Association of China, a self-regulating industry body, is demanding more disclosure from major quant players about their operations. Yin Tianyuan, the Shanghai-based head of research at data provider Suntime Information, says some quant managers are voluntarily taking the heat out of their trading. “The largest quants with 100 billion yuan in China have almost all lowered the frequency of their algorithm to ‘medium high’,” she said.
(Editing by Vidya Ranganathan and Sam Holmes)