The Impact of ESG Report Readability on the Risk of Greenwashing

Purpose: While prior research has focused on ESG disclosure content, little attention was paid to how textual complexity may facilitate deceptive practices. Using legitimacy and information asymmetry theories, this study investigates whether firms strategically reduce ESG report readability to mask greenwashing. Research design, data, and methodology: Analyzing 8,742 firm-year observations from 652 Chinese manufacturing companies with Class A shares listed from 2009 to 2023, readability was measured via an innovative large language model (DeepSeek R1). Results: A significant negative relationship was indicated between readability and greenwashing. For every one-unit increase in readability, the greenwashing indicator decreased by 0.331 units. Findings are robust across alternative models (ChatGPT-4o, QwQ- 32B, DeepSeek v3, fog_index) and under multiple endogeneity checks. Conclusions: This research contributes to the literature by empirically demonstrating textual complexity as a strategic greenwashing tool, introducing a superior AI-based readability measure, and providing practical policy implications, highlighting readability standards as an effective regulatory measure against greenwashing.