A multi-dimensional textual framework for detecting greenwashing in sustainability reporting

Drawing on impression management theory, this study investigates the presence of potential greenwashing in sustainability reports (SRs) by examining three textual dimensions: readability, sentiment, and boilerplate content. A novel three-dimensional categorization framework was developed to capture companies’ sustainability communication strategies. Using text mining and natural language processing techniques, we analyzed 105 English- language SRs published by companies listed on the BIST Sustainability Index between 2021 and 2023, comprising over 250,000 sentences. The findings reveal a general trend toward more readable and positive narratives; however, in some cases, this was accompanied by increased use of boilerplate language, suggesting impression management practices. Notably, a small group of companies consistently exhibited a high-risk communication profile indicative of potential greenwashing. This study contributes to the literature by integrating three complementary textual indicators into a unified model for detecting greenwashing behaviors. The results provide practical implications for regulators and stakeholders seeking to enhance the credibility and transparency of corporate sustainability reporting.