Portfolio optimization methods for highly volatile assets: challenges and solutions

Portfolio optimization is a crucial aspect of managing finances for institutions. It involves deciding how to distribute wealth among different assets. Traditional methods of portfolio optimization have limitations, especially when there are many assets or extreme values, which can lead to unstable outcomes. Addressing these challenges, a new study introduces a promising approach called the minCluster portfolio method. It is specifically designed for highly volatile markets, including those involving cryptocurrencies.

“The minCluster method combines different techniques to achieve the best allocation of assets in a portfolio. It uses a unique way of analyzing the relationships between assets and considers the downside risk,” shared Emmanuel Jordy Menvouta, first author of the study. “This allows it to find the best combination of assets, even in highly unpredictable markets. The methodology used in this approach has the potential to be applied to other areas beyond portfolio optimization.”

The research findings show that the minCluster portfolio optimization method is effective in constructing robust portfolios in highly volatile markets. This has significant implications for investors and portfolio managers who want to optimize their portfolios in challenging market conditions. Notably, the minCluster method consistently outperforms traditional approaches in tests conducted outside the sample. By using robust techniques and considering downside risk, this approach has the potential to improve portfolio performance and reduce the impact of losses.

Menvouta and his co-authors published their findings in a KeAi journal, The Journal of Finance and Data Science.

"This is the first paper to combine ideas from robust statistics, machine learning, and portfolio optimization. It opens up new avenues for building portfolios that can handle extreme situations and perform well in highly volatile markets,” added Menvouta.

The team believes that further research may explore how the minCluster method performs with different datasets and market conditions.

###

Contact the author:

• Emmanuel Jordy Menvouta, KU Leuven, Department of Mathematics, Section of Statistics and Data Science, Leuven, Belgium, emmanueljordy.menvoutanpkwele@kuleuven.be
• Sven Serneels, Department of Mathematics, University of Antwerp, B-2020 Antwerp, Belgium, sven.serneels@uantwerpen.be
• Tim Verdonck, Department of Mathematics, University of Antwerp - imec, B-2020 Antwerp, Belgium, tim.verdonck@uantwerpen.be

The publisher KeAi was established by Elsevier and China Science Publishing & Media Ltd to unfold quality research globally. In 2013, our focus shifted to open access publishing. We now proudly publish more than 100 world-class, open access, English language journals, spanning all scientific disciplines. Many of these are titles we publish in partnership with prestigious societies and academic institutions, such as the National Natural Science Foundation of China (NSFC).