Stockholm (HedgeNordic) – There is one common perception that trend-followers are long volatility. Not every type of volatility, however, is necessarily suitable for trend-followers, as volatility can manifest itself in many shapes and forms. Mandatum Life Managed Futures considers this changing nature of volatility by using machine learning algorithms that select the right combination of trading models for a given environment. As a systematic fund, Mandatum Life Managed Futures allocates risk to various models including trend following.
The sudden and wild coronavirus-driven market sell-off in March may have caught Mandatum’s artificial intelligence-assisted systematic vehicle by surprise, but the system quickly rebooted and started to capture trends in the volatile markets of the subsequent four months. Launched by Finnish life insurer Mandatum Life in late December last year, Mandatum Life Managed Futures was the best-performing member of the NHX CTA sub-index over the past four months with a return of 10.5 percent. The fund, powered by machine learning algorithms designed by artificial intelligence firm PROWLER.io, is now up 4.5 percent since launching on December 30 last year through August 11.
“The performance is truly excellent taking into account what has happened in the markets.”
“The performance is truly excellent taking into account what has happened in the markets,” Ville Rantanen (pictured), portfolio manager at Mandatum Life Insurance, comments on the fund’s performance. “We are especially pleased that the results also align with our historical simulations and backtests,” Rantanen continues. Before launching Mandatum Life Managed Futures, the team at Mandatum Life simulated how the artificial intelligence algorithms behaved historically during significant market turbulence. “Our results showed that AI came out as a winner during the 1987 stock market crash, the early 90’s recession, the tech bubble burst in 2000 and the financial crisis of 2008/2009,” says Rantanen.
The Mandatum Life team also “paper traded” the eight months before going live at the beginning of 2020, with the strategy outperforming the SG CTA index by 4.5 percent. More importantly, the “correlation between the AI simulation and “paper trading” was 0.98 during that period, validating the ability of the system to simulate real world activity including costs and other operative factors,” emphasizes Rantanen.
“The fund changes from running shorter-term trading models to longer-term ones depending on which types of strategies the machine learning algorithms believe are more suitable to the current environment.”
Mandatum Life Managed Futures uses artificial intelligence developed for Mandatum by a Cambridge-based specialized AI company to “pick and choose the right strategy depending on the market environment,” as previously explained by Lauri Vaittinen, who contributed to the development of the strategy. “The fund changes from running shorter-term trading models to longer-term ones depending on which types of strategies the machine learning algorithms believe are more suitable to the current environment.” Instead of employing a fixed trend-following strategy that generates “returns that are zig-zagging around zero over the years,” Mandatum’s systematic vehicle relies on machine learning to select the right combination of trading models for a given environment.
“During the first half of 2020, the AI has already shown its ability to mitigate short-term fluctuations and to help portfolio management make investment decisions that maximize returns from the mid- and long-term trends.”
“During the first half of 2020, the AI has already shown its ability to mitigate short-term fluctuations and to help portfolio management make investment decisions that maximize returns from the mid- and long-term trends,” says Rantanen. The performance of Mandatum Life Managed Futures “has been very good both in absolute and in relative terms, when comparing against the SG CTA Index,” he continues. “The Fund has returned 4.6 percent since inception, while CTA Index still shows a negative return of -0.3 percent for the same time period.” Rantanen goes on to say that “this return has come with an attractive risk-return Sharpe ratio of one, despite extreme market volatility.”