Financial Market Challenges and ML-Supported Asset Allocation Financial Market Challenges and ML-Supported Asset Allocation
Editor’s note: Peter Schwendner, PhD is a speaker for ODSC Europe this June. Be sure to check out his talk, “ML... Financial Market Challenges and ML-Supported Asset Allocation

Editor’s note: Peter Schwendner, PhD is a speaker for ODSC Europe this June. Be sure to check out his talk, “ML Applications in Asset Allocation and Portfolio Management,” there!

The year 2022 presented two significant turnarounds for tech: the first one is the immediate public visibility of generative AI due to ChatGPT. The second one is the rise in interest rates after several decades of falling interest rates, leading to stalling equity markets. Both drivers challenge established business models relying on hardware and leverage from different angles.

Low-interest rates of the last decades have pushed many investors into illiquid asset classes like private markets and infrastructure. An important argument for illiquid assets is that its risk premia do not come together with daily visible price volatility. In contrast to that, the visible volatility in liquid markets needs to be managed on a portfolio level. Therefore, in institutional asset management, dynamic multi-asset allocation heuristics like “risk parity” are used widely for liquid portfolios. A significant driver for the demand was the successful performance of risk parity during the Global Financial Crisis and during the European Sovereign Debt Crisis, as the major government bonds could diversify equity drawdowns. 

However, changing correlations can be a challenge for this type of portfolio allocation technique. For example, rising interest rates and falling equities already in 2013 and again in 2020 and 2022 led to drawdowns of risk parity schemes. In 2023-Q1, we even saw failing banks like SVB simply because of investments in “safe” treasury bonds. 

In the last years, several machine learning innovations have been introduced to improve the robustness of asset allocation with hierarchical clustering and seriation-based approaches, to improve the transparency of these heuristics with explainable AI, and to generate synthetic correlations and correlated market returns to enhance the coverage of backtests and scenario analysis beyond the historical paths. Together, these innovations offer a consistent pipeline for a better understanding of rule-based dynamic portfolio allocation strategies. My talk “ML Applications in Asset Allocation and Portfolio Management” on June 15 at ODSC Europe in London reviews ML applications to asset allocation and puts them into the context of the current market situation.

Peter Schwendner leads the Institute of Wealth & Asset Management at Zurich University of Applied Sciences, School of Management and Law, Switzerland. His interests are financial markets, asset management and machine learning applications. With the European Stability Mechanism (ESM), he has been developing analytics for primary and secondary bond markets and tools for optimizing the issuance process. Currently, he is working on the BRIDGE Discovery project “Spatial sustainable finance: Satellite-based ratings of company footprints in biodiversity and water”. Within the European COST Action «Fintech and AI in Finance», he leads the working group «Transparency into Investment Product Performance for Clients».

ODSC Community

The Open Data Science community is passionate and diverse, and we always welcome contributions from data science professionals! All of the articles under this profile are from our community, with individual authors mentioned in the text itself.