ESG and systemic risk in financial networks: High- vs. low-rated portfolios in Europe

We examine ESG-integrated portfolio strategies using a dynamic, network-based approach applied to the European financial sector. Using the STOXX Europe 600 index, we construct two minimum-variance portfolios—one of high-ESG-rated assets and the other of low-ESG-rated assets—based on quarterly networks of financial institutions that reflect systemic relevance. The methodology combines DCC-GARCH correlations with heteroscedasticity-robust Granger causality to capture time-varying co-movement and directional dependencies. From these relationships, we derive the core financial network, filter it with a minimum spanning tree, and apply spectral clustering to identify structurally similar groups of assets. Portfolios are formed by selecting cluster representatives according to ESG scores. Results show that while low-ESG portfolios outperformed on average during 2012–2022, high-ESG portfolios delivered superior cumulative returns and risk-adjusted performance in 2023, coinciding with the EU Corporate Sustainability Reporting Directive. Our findings highlight the potential of ESG criteria to enhance resilience and performance within a network-aware framework.