Helpful and supportive for the human being - artificial intelligence opens the field of vision. Photo: mevans Source: iStock

There is a bigger picture

Author: Stefan Riße and Timo Lüllau-Mortensen

Some relationships are too complex for humans to see - because our minds tend to run too much along traditional patterns. Artificial intelligence (AI) helps companies to think outside the box. In the interview with ChampionsNews, Stefan Riße, capital markets strategist at ACATIS Investment Kapitalverwaltungsgesellschaft, explains how artificial intelligence is used for stock selection, for instance in the ACATIS Global Value Total Return (ISIN: DE000A1JGBX4).

Stefan Riße, capital markets strategist at ACATIS Investment Kapitalverwaltungsgesellschaft Source: ACATIS

ChampionsNews: Mr Riße, you have been relying on artificial intelligence since October 2016 to preselect candidates for the ACATIS Global Value Total Return (ISIN: DE000A1JGBX4) . According to which criteria does the Convolution Network select stocks that meet the criteria for potential portfolio inclusion?

Riße: In a first step, stocks are selected by an AI-based proprietary ACATIS model that analyses the companies in our ACATIS fundamentals database. It relies on 55 fundamental parameters to screen the about 2500 strong universe for stocks that should outperform over the short, medium and long term, usually narrowing the universe down to about one third of its original size in the process. The criteria include very traditional ratios such as dividend yield, sales, earnings or book value. In a next step, we put the selected stocks through our value screenings that result in 50 stocks with particularly positive value characteristics and valuations. Our portfolio managers will subsequently analyse these in more detail from a qualitative point of view. Stocks are equally weighted and held for between one and three years on average before being replaced.

More importantly, however, machine learning allows us to identify relationships between parameters within companies and sectors, across sectors and over time.

Stefan Riße

ChampionsNews: Where do you perceive advantages in the use of artificial intelligence? Does the pre-selection process identify stocks that would otherwise be overlooked by your active investment approach?

Riße: Firstly, the use of artificial intelligence enables us to screen a much larger amount of data than we could hope to cope with in manual processes. More importantly, however, machine learning allows us to identify relationships between parameters within companies and sectors, across sectors and over time. Artificial intelligence analyses relationships we humans wouldn’t even consider due to their apparent complexity. They do exist, however. Our human mindset relies on experience that makes us think in patterns we have learned to trust; machines, on the other hand, are unprejudiced. That opens up entirely new vistas.

ChampionsNews: Which stocks have recently proven particularly successful for the portfolio and why did you select them?

Riße: Swedish Match AB, a Stockholm-based industrial with a long heritage that focuses on tobacco and match production, has done very well recently. Founded in 1917 by Ivar Kreuger, later to be dubbed the “King of Matches”, it came to life as „Svenska Tändsticksaktiebolaget“. In a tragic twist of fate, Ivar Kreuger later became world-famous when he took his own life in Paris shortly before his trust became insolvent. The company was saved eventually and exists to this day. Another positive contribution came from Microsoft Corporation that has made it back into the top league, among other things thanks to the cloud business. Since all stocks are selected according to quantitative fundamental criteria, the reason for their inclusion in the portfolio is rather banal. The AI and subsequently the portfolio manager pick them based on their attractive characteristics and valuation.

ChampionsNews: How variable is your exposure?

Riße: The fund’s exposure is managed based on quantitative criteria and relies on two performance drivers: alpha from stock selection and beta from the allocation decisions that reflect our bottom-up and top-down views. The equity exposure can range between zero and 110 percent, which means that the fund is highly flexible.


Author: Stefan Riße and Timo Lüllau-Mortensen
Date of issue: 12/7/2018