Kvartil, an asset manager based in Stockholm, has created the Investmentbolag+Calculus fund, which will use machine learning tools to try to generate a combination of greater returns and lower risk than an index consisting of listed investment companies such as Investor AB, Kinnevik AB, Lundbergforetagen AB and Industrivarden AB.

The trick is to capture opportunities that arise due to changes in the difference between such firms’ market capitalization and the net asset value of their holdings, Kvartil Chief Executive Officer Torbjorn Olofsson said in an
interview.

“This discount varies much over time and is a risk to any investor in investment companies,” Olofsson said. “We have created rule-based strategies that aim to exploit this variation in order to generate alpha.”

Since the discount can be observed, it offers a more or less objective measure of which investment companies are cheap and which ones are expensive, compared with alternative investment, Olofsson said. That makes investment companies unique from an analytical point of view, he said.

Kvartil estimates that the new fund’s model can generate about 2 percent annually in excess returns compared with a more traditional portfolio of investment companies. It generated an annualized return of 19.1 percent at 19.5 percent volatility over the period 2003-2019.