Decoding Greenwashing: Llm Insights into Corporate Narrative

Abstract

This study focuses on the role of narrative in corporate disclosures and proposes a novel framework for systematically evaluating the characteristics of the corporate disclosures released by firms accused of greenwashing. Greenwashing refers to the practice of presenting sustainability-related information in a manner that misrepresents a firm's actual practices; such practices undermine corporate credibility and influence stakeholder decision-making. We use large language models (LLMs) to analyze 3,108 news articles from English-language media sources published between 2016 and 2022 to identify firms associated with greenwashing. Additionally, we introduce a methodology for quantifying corporate narrative quality, employing LLMs to compute readability scores for 2,261 sustainability reports. Our regression analysis of readability scores in relation to greenwashing allegations indicates that the reports of firms that are not implicated in greenwashing incidents tend to receive higher readability scores. This finding suggests that companies with clearer and more logically structured disclosures are less likely to engage in greenwashing.

Keywords: greenwashing, ESG, Large language model, Narrative, readability