Management communication significantly influences investor perceptions, information asymmetry, and market behavior. Traditional sentiment-based approaches applied to corporate disclosures often overestimate positivity and fail to discriminate across firms due to optimistic language biases and limited sensitivity. This study proposes a Strict Management Communication Index (MCI), an advanced NLP-driven scoring model designed to quantify managerial tone, clarity, uncertainty, and sentiment consistency across multiple corporate documents. The framework extracts text from annual reports, earnings call transcripts, and investor presentations using high accuracy PyMuPDF extraction and evaluates sentiment using the domain specific FinBERT model. Key innovations include (i) a Strict Sentiment Score penalizing neutral-heavy communication, (ii) TF-IDF inspired normalization of optimism, risk, and uncertainty, (iii) a readability penalty based on the Gunning Fog Index, and (iv) uncertainty amplification to penalize evasive communication. We evaluate the model on BSE Sensex 30 companies and show that the Strict MCI produces a significantly wider and more realistic distribution compared to conventional sentiment scores. The MCI can act as a forward-looking soft-information factor in equity selection and portfolio optimization.
BHARAD et al. (Sun,) studied this question.