AI hedge fund Voleon suffers in choppy markets - Financial Times

doro.indah.link

Voleon Group, one of the world’s biggest and best-performing artificial intelligence hedge funds, has suffered large losses after being hit by choppy markets, making it one of a number of computer-driven fund managers to struggle this year.

The California-based fund, which manages about $6bn in investor assets and which was set up by former DE Shaw traders Michael Kharitonov and Jon McAuliffe, has lost about 9 per cent this year in its flagship Investors fund, said a person who had seen the numbers. Its more recently-launched Institutional Strategies fund has lost 1.5 per cent this year.

Voleon declined to comment.

The losses are surprising given that Voleon has been a strong performer in recent years. Its Investors fund had made double-digit gains in 2018’s choppy markets, when the S&P 500 lost 4.4 per cent after sell-offs at the start and end of the year. The fund also made money in last year’s strongly-rising markets.

However, like many computer-driven funds, Voleon was caught out by this year’s market turmoil. That had left the Investors fund down by double digits earlier this year, and the Institutional Strategies fund down single digits, say people familiar with its returns.

In August the main fund clawed back 2.5 per cent, while the Institutional Strategies fund gained 2 per cent.

High-profile quant funds such as Jim Simons’s Renaissance Technologies and David Harding’s Winton Group have also suffered losses in some funds this year as algorithms have struggled to cope with extreme market moves.

Andrew Beer, managing member at US investment firm Dynamic Beta Investments, said some machine learning funds “learned to buy dips” in the market between 2015 and 2019, “so they caught the proverbial falling knife in the first quarter”.

He added: “To make matters worse, some managers overrode the models at the bottom and missed the bounce.”

Voleon is unusual in that, while many quantitative fund firms use a small element of machine learning in their investment process, very few focus solely on this approach. Machine learning involves letting algorithms learn and extrapolate rules from the data, rather than a human setting the rules an algorithm should follow. The firm trades a strategy known as statistical arbitrage, which involves betting that short-term discrepancies in prices will revert to a mean.

The firm has been bulking up its research and development effort, hiring about 40 people since March this year. It now has about 200 employees.

Voleon’s Investors fund has largely made money in recent years, although it lost money in 2016. Its Institutional Strategies fund, which was launched in 2016 and which trades less than the main fund, also lost money that year but made double-digit gains the following year, according to an investor letter seen by the Financial Times.

AI hedge funds have gained 14.5 per cent on average this year, according to Eurekahedge’s AI Hedge Fund index. However, that was largely due to an exceptionally strong August in which funds were up 12.2 per cent.

laurence.fletcher@ft.com

September 07, 2020 at 11:01AM
https://ift.tt/3bAD3xQ

AI hedge fund Voleon suffers in choppy markets - Financial Times

https://ift.tt/2APQ0pp