A Multi-factor Climate Risk Index (CRI) for fossil fuels and ESG investments: A jar full of CRI-kets

This paper introduces a novel multi-factor Climate Risk Index (CRI) that captures the multifaceted nature of climate change risks and their impact on economic and financial markets, especially energy and ESG investments. This study fills a research gap by creating a multi-factor CRI that combines nature-based observations, news sentiment, and policy uncertainty indicators. It uses a new two-stage process to capture climate risk's nonlinear complexities, unlike existing literature that focuses mainly on single metrics related to physical events or news sentiment. We build CRIs incorporating climate risk components from established indices using three key sources: climate change news sentiment, physical climate impacts, and policy developments. Through multi-stage modelling—including PCA for initial extraction, feature engineering, and consolidation; t-SNE for nonlinear dimension reduction; and PLSR to integrate market response variables and create tailored CRIs for fossil fuels and ESG investments—we achieve comprehensive climate risk assessment. We also develop a "super-CRI" by applying a meta-combination technique, the Wavelets Ensemble (METS 3.0), to synthesize diverse climate risk indicators into a single comprehensive measure. Our proposed multi-factor CRI represents a significant step forward in understanding and pricing climate risk in financial assets, outperforming single-factor indices as a predictor of ESG and fossil fuels risk premia.