Executive Summary
This study introduces a Hierarchical Variable Selection (HVS) algorithm designed to identify the ESG variables most strongly associated with corporate financial risk. Using raw data from the London Stock Exchange Group (LSEG), the authors demonstrate that HVS achieves higher explanatory power and greater precision than traditional ESG scoring methods and statistical models. The results show that while ESG variables have limited predictive value for returns, they are highly informative for understanding corporate risk, especially sector- and size-specific variations in volatility. The HVS framework provides a data-driven and transparent approach to identifying material ESG risk factors, offering valuable insights for investors, regulators, and analysts seeking to enhance risk modeling, portfolio resilience, and sustainable finance practices.
A New Approach to ESG Risk Analysis
The paper “Identifying Risk Variables from ESG Raw Data Using a Hierarchical Variable Selection Algorithm” (Chen, Feinstein, & Florescu, 2025) presents a method for extracting ESG risk indicators directly from unaggregated data. The authors develop the HVS algorithm to utilize the hierarchical structure of ESG datasets, pinpointing the raw variables most closely linked to corporate financial risk as measured by the logarithmic volatility of returns. Traditional ESG ratings condense hundreds of individual metrics into a single score, often obscuring key risk drivers and resulting in inconsistent outcomes across different rating agencies. The HVS approach instead examines raw ESG data from the LSEG database to identify the indicators that most effectively explain financial volatility, enhancing both accuracy and interpretability in ESG risk evaluation.
Model Performance and Key Findings
The findings demonstrate that HVS outperforms well-known variable selection methods, such as PCA, Lasso, and stepwise regression, by providing greater explanatory power with fewer variables. The algorithm suggests that ESG indicators exhibit a weak correlation with stock returns but a strong, statistically significant relationship with risk. Notably, the relevant ESG variables vary significantly across sectors. In the energy industry, for example, social metrics related to employment quality and community relations account for more volatility than environmental factors, while in finance and utilities, governance factors are more influential. Firm size also affects the results: large-cap companies exhibit stronger governance-related risk signals, whereas small-cap firms are more vulnerable to environmental and innovation-related variables.
Integrating ESG with Traditional Financial Models
The study further combines HVS-selected ESG factors with the traditional Fama-French three-factor model, demonstrating that adding ESG data significantly enhances the model’s ability to explain risk across all sectors. This integrated approach demonstrates that ESG variables provide additional insights beyond traditional financial factors, improving the understanding of volatility and downside risk, rather than focusing solely on short-term returns. By directly linking sustainability metrics to financial risk measures, the research connects corporate responsibility data with quantitative finance, demonstrating that ESG information can significantly enhance risk management and decision-making.
Implications for Risk Modeling and Sustainable Finance
Overall, the authors conclude that ESG data at the raw-variable level contains valuable information about financial risk, even if it offers little for return prediction. The HVS algorithm provides a robust, transparent, and straightforward approach for identifying sector-specific ESG risk drivers across firms of varying sizes. Its use enhances risk modeling accuracy, facilitates more reliable sustainability reporting, and provides a stronger empirical basis for integrating ESG analytics into financial systems. In a rapidly evolving investment environment, the HVS framework helps connect sustainability research with financial practice by demonstrating how specific ESG factors can be translated into measurable indicators of corporate resilience.
Frequently Asked Questions (FAQs)
- What is the Hierarchical Variable Selection (HVS) algorithm? The Hierarchical Variable Selection (HVS) algorithm is an AI-powered, data-driven method used to identify the most relevant Environmental, Social, and Governance (ESG) variables linked to corporate financial risk. By leveraging hierarchical data structures and advanced machine learning, HVS pinpoints material ESG risk factors with greater accuracy and transparency than traditional ESG scoring systems.
- How does HVS improve ESG risk modeling compared to traditional scoring methods? Traditional ESG scores condense hundreds of indicators into a single, opaque rating, often obscuring the real sources of risk. The HVS algorithm uses artificial intelligence to analyze raw ESG data directly, revealing which variables drive financial volatility. This improves explainability, precision, and trust in ESG analytics for investors, regulators, and asset managers seeking credible, data-backed insights.
- What are the key findings from the HVS ESG risk study? The study shows that while ESG variables have limited predictive power for short-term returns, they are highly informative for corporate risk and volatility. Results from the London Stock Exchange Group (LSEG) dataset reveal that relevant ESG factors vary by sector: social indicators dominate in energy, governance in finance and utilities, and environmental innovation in smaller firms—offering clear, AI-enhanced visibility into sector-specific vulnerabilities.
- How does the HVS algorithm integrate with financial models? The HVS-selected ESG variables were incorporated into the traditional Fama-French three-factor model, significantly improving its ability to explain volatility and downside risk. This AI-assisted fusion of ESG data and financial modeling enables investors to understand resilience better, enhance portfolio performance, and align capital with long-term sustainability outcomes.
- Why is the HVS approach necessary for sustainable finance and risk intelligence? The HVS framework represents a breakthrough in ESG risk intelligence by translating raw sustainability data into measurable financial insights. It allows analysts, asset managers, and regulators to identify material ESG factors, quantify their economic impact, and strengthen risk modeling. This AI-driven, transparent methodology supports sustainable finance, resilience planning, and evidence-based investment decisions.
(Source: Chen, Z., Feinstein, Z., & Florescu, I. (2025). Identifying risk variables from ESG raw data using a hierarchical variable selection algorithm. Stevens Institute of Technology. https://arxiv.org/abs/2508.18679.)
Ready to get started? To learn how ClimaTwin can help you assess the physical and financial impacts of future weather and climate extremes on your infrastructure assets, capital programs, and investment portfolio, please visit www.climatwin.com today.
© 2025 ClimaTwin Corp. All rights reserved worldwide.
ClimaTwin® is a registered trademark of ClimaTwin Corp. The ClimaTwin logos, ClimaTwin Solutions™, Climate Risk Intelligence™, Climate Business Intelligence™, Future-proofing assets today for tomorrow’s climate extremes™ are trademarks of ClimaTwin Corp. All trademarks, service marks, and logos are protected by applicable laws and international treaties, and may not be used without prior written permission of ClimaTwin Corp.
###