Unpacking banks’ response to societal expectations: An NLP analysis of European banks’ discussion of climate change

Artikel i vetenskaplig tidskrift, 2026

We employ Natural Language Processing (NLP) to analyze how climate change is discussed in the annual sustainability reports of the 35 largest EU banks (2015–2021), assessing alignment with four societal expectations: decarbonizing financial products and services, addressing climate-related risks, reducing operational emissions, and enhancing transparency. These expectations stem from governments, regulators, and civil society. Analyzing over 1.5 million statements, we find that about 7% of content pertains to climate change. Banks increasingly focus on decarbonizing consumer products and their own operations, but devote less attention to financed emissions, transition risks, and concrete commitments. Our study contributes to the application of NLP in climate finance by qualitatively interpreting how banks, in their own words, engage with societal expectations around the climate transition. This complements quantitative studies by contextualizing disclosure patterns and highlighting reporting gaps and under-emphasized but material issues, offering insights that can inform policymakers in designing disclosure requirements.