Stockholm (HedgeNordic) – Austrian CTA manager SMN is clocking in at an impressive gain of 85 percent with their SMN Diversified Futures Fund over the rolling 36-month period. This comes after a tough decade for the strategy with a new maximum drawdown of 39 percent since the fund´s inception in 1996. Persistence and the ability to sit out such periods seem to be key to the long-term success of a systematic manager. But surely, the heat is on and the pressure goes up in such periods.
“Investor pressure was definitely high. Actually, going back to maybe 2018/2019, our investors were telling us we may be the last ones believing in the good, old trend following story and frequently suggested to make fundamental changes to the investment strategy, adapting it to the so called new normal,” SMN’s managing director Gernot Heitzinger recalls.
Assets under management decreased for the Austrian manager during this period and two of the largest investors redeemed their allocations in early 2021, just when performance started to gain momentum again. Christian Mayer and Michael Neubauer, two of SMN´s founding partners, who are still active in daily business, have experienced this kind of investor behavior before. “In our 26 years running the strategy, we have seen this before. It was similar in 2007 and 2013 until performance kicked in. Investors throw in the towel at the most unfortunate time. These experiences taught us not to overreact and panic, not to change the basics of our trading philosophy, strategies and fundamental beliefs.”
In the tough times of a long negative performance cycle, there may be a fine line between being heroically persistent, stubborn, inflexible or simply wrong. Being systematic and rules-based can help evaluating decisions in the rear-view mirror.
“Changing too much in periods of stress is probably not the best timing, as you adapt to isolated situations, causing more harm than good in the long run,” Joseph Waldstein, who is leading SMN´s quant research team, emphasizes the importance of having a solid foundation in ones’ beliefs and understanding of markets and being persistent rather than short-term opportunistic.
Same Model Since 1996
“We have models unchanged since 1996, others have been slightly adapted while the fundamental logic is still the same. We did a lot of research and implementations in areas like portfolio diversification and risk management. Overlays for improving management of changing market risk regimes and cluster risks limitation techniques are two examples. These innovations had a larger effect than changing components such as time frames of trading, or trend detecting methodologies,” says Heitzinger.
“We have models unchanged since 1996, others have been slightly adapted while the fundamental logic is still the same.”
“For us, diversification and identifying diversifiers always have been key elements,” Neubauer is convinced. “We started trading a portfolio of 50 markets in 1996 and ended up with around 80 a decade later.” A significant increase in the number of markets traded came with the introduction of synthetic markets in 2008. These are combinations of different contracts, such as spreads between markets, calendar spreads and baskets of different instruments.
During the financial crisis, investors were cashing in on managed futures funds who had performed exceptionally well, remained liquid and offered protection from tumbling markets. “Later Investors returned to the asset class but funds were allocated to a small number of very large advisors, rather than smaller ones such as a boutique based in Vienna,” Heitzinger recalls.
Small and Nimble, Yet True to Trend
“The advantage for us being small and nimble was the ability of doing something different,” says Waldstein. SMN started diving into a concept Waldstein refers to as “Structural Alpha Trend”, meaning trading niche markets being off the radar for large traders. “The portfolio effects as a result from trading smaller commodity markets such as South African grains, milk, cheese and many others are impressive due to the low correlation to traditional portfolios” Waldstein explains. Trading returns in these markets show a relatively low correlation of 0.2 or 0.25 to a traditional trend following portfolio, compared to more traditional markets averaging correlations between 0.7 to 0.8. Adding such niche markets allows SMN to identify exploitable trading opportunities in more than 300 instruments. This naturally comes at the expense of limited capacity which SMN commits in order to offer unique diversification for its investors.
“The advantage for us being small and nimble was the ability of doing something different.”
Waldstein describes how SMN is constantly challenging existing models or developing new ideas. “A high percentage of our research efforts is not implemented into our trading systems. While this seems to be disappointing, understanding what works and what does not is key for us, assuring the validity of our approach and our level of confidence.”
Performance and marketing pressure tends to influence the CTA industry heavily, provoking possible style shifts. One notable example may be the move from trend following to multi-strategy, multi-style approaches. While that seems to work for a handful of CTA companies we have seen many more failing with such style shifts. SMN has resisted such temptations, remaining dedicated to a pure trend following strategy.
More recent evolutions such as machine learning and artificial intelligence are features SMN may use in their research and development activities, but do not see them being implemented in active trading models. “We are very cautious gearing towards a trading methodology which is self-learning. We are concerned that such tools may take too much emphasis on the most recent past. At least there is a big style-drift risk, without even noticing it. Furthermore adapting for shorter time intervals has not really proven to be helpful in our history,” Waldstein is skeptical.
“We are very cautious gearing towards a trading methodology which is self-learning. We are concerned that such tools may take too much emphasis on the most recent past.”
Other new attributes such as alternative data is also something SMN researched but rejected so far. Neubauer explains: “Alternative data may add to your systems if you really believe in market inefficiency and fundamental advantage through additional information. We always believed in the behavioral nature of price driven trend following, so we doubt that such fundamental inputs will consistently add value to our approach.”
Optimizing models by results of back testing is very dangerous. “During the period from 2009 to 2021 we would have better stopped trading any markets on the short side as a result of the asset bubbles caused by central banks. But so far in 2022 much of our return comes from short positions in fixed income.” Waldstein highlights.
A long-term view and a strong conviction to the strengths of the strategy is crucial. Panicking and changing key features because of disappointing results is emotional. One of the biggest advantages of technical trading is to overcome emotions.
While some of the largest CTAs employ large numbers of researchers, such resources are typically not available to boutiques the size of SMN, currently managing around $200 million with the help of eight staff involved in research and portfolio management. Neubauer doubts the benefits of engaging an army of researchers. “Sure, it may be helpful to have individual projects done quicker, but on the other hand a large team of researchers will likely produce too many conflicting ideas in day-to-day operations. There could be a lot of noise while being at risk of becoming overcomplex. Getting the big picture right is likely the most important part in your modelling. Employing an army of PhDs is not in the cards for a boutique like ours, but we used to work with a much larger research department years ago. For us it does not provide the big advantage the numbers would suggest.”
It seems, the temptation to overengineer and overoptimize at times may lure you to fix something that was in fact never really broken.
This article features in HedgeNordic’s “Systematic Strategies” publication.