A text mining analysis of European banks' and insurers' disclosures on climate‐related risks

Anna Kraus
DOI: https://doi.org/10.1111/rmir.12268
2024-05-12
Risk Management and Insurance Review
Abstract:This paper empirically analyzes sustainability‐related disclosures of listed European banks and insurance companies by focusing on transition risks, physical risks as well as stranded assets as a possible consequence of transition and physical risks. We employ a text mining analysis and apply a comprehensive keyword catalog to 142 firms. The dataset used for the analysis comprises 852 annual and 551 sustainability reports, covering a sample period from 2017 to 2022. Our results show that disclosure on climate‐related risks considerably increases over the examined period, but that reporting on stranded assets is still in early stages. Moreover, European insurers report more on climate‐related risks relative to the number of analyzed tokens than banks. Finally, statistical evidence suggests that larger banks and insurers with above‐median market capitalization exhibit significantly more word hits on climate‐related risks as compared to smaller firms.
What problem does this paper attempt to address?