Research on the Impact of ESG text on Corporate Value Based on Implicit Semantic Sentiment Analysis

Amid the global surge in interest in ESG disclosure, prior studies have predominantly focused on structured ESG information while neglecting implicit semantic sentiments - the latent subtle emotions and tone embedded within textual narratives. To address this gap, we propose a Knowledge Graph Driven Sentiment (KGDS) model, which employs a hybrid deep learning architecture comprising text-processing neural net works and a graph convolutional network to quantify the implicit sentiments embedded in ESG disclosure documents. Using a sample of 500 firms spanning the period from 2017 to 2022, our empirical analysis investigates the impact of these implicit sentiments on corporate market value, proxied by Tobin’s Q. Our analysis reveals three key findings. First, positive implicit sentiment positively affects Tobin’s Q, while negative implicit sentiment has the opposite effect, with these sentiment effects often surpassing the valuation impacts of explicit ESG performance and firm characteristics. Second, the sentiment effects exhibit pronounced industry heterogeneity, with firms in the financial and energy sectors being particularly sensitive to sentiment signals. Third, our results provide no evidence of an “excessive positivity saturation effect”. SHAP-based explain-ability analysis shows that implicit sentiment ranks as the second most influential factor in determining market valuation, surpassed only by firm size. Theoretically, this study introduces a novel analytical framework: “textual sentiment $\rightarrow$ investor perception $\rightarrow$ firm action $\rightarrow$ value feedback”. This framework offers a distinct perspective on ESG analysis, shifting the focus away from purely quantitative ESG metrics to incorporate qualitative textual elements. Practically, our findings suggest that securities authorities should incorporate advanced textual sentiment analysis into ESG disclosure guidelines to enhance the detection of greenwashing practices, and listed companies should strategically tailor the narrative style of their ESG reports.