Beyond polarity: How ESG sentiment influences idiosyncratic volatility in the Turkish stock market

Abstract

This study investigates the influence of Environmental, Social, and Governance (ESG) sentiment in corporate disclosures on idiosyncratic volatility (IVOL) in the Turkish stock market. Using FinBERT-ESG, a language model specifically designed for financial and ESG-related texts, we construct four novel indices: the Positive ESG Index (PESGIN), capturing positive ESG sentiment; the Negative ESG Index (NESGIN), representing adverse ESG sentiment; the Balanced Polarity Index (BPI), measuring the overall balance between positive and negative sentiment; and the Amplified Negative Polarity Index (ANPI), which emphasizes the intensity of negative sentiment. By employing a system-GMM approach, which effectively addresses endogeneity concerns common in finance, we find that PESGI is negatively associated with IVOL, suggesting that transparent and optimistic ESG communication reduces firm-specific risk. Conversely, ANPI positively correlates with IVOL, supporting the overreaction hypothesis and highlighting elevated investor sensitivity to adverse ESG disclosures. These findings underscore the complex interplay between ESG sentiment and investor behaviour, offering valuable insights for enhancing risk assessment and guiding investment strategies.