Sentiment in Sustainability: Does mood in ESG Disclosure Associate with Financial Performance in the Real Estate Sector?

Traditional ESG ratings, which primarily rely on structured data and qualitative assessments, often fail to capture the nuanced nature of ESG-related communications. Sentiment analysis using computational linguistics approaches offers a valuable complementary approach by systematically capturing the tone of ESG disclosures and media coverage, providing deeper insights into corporate sustainability commitments and stakeholder perceptions. This study adopts a novel approach using textual information set and investigates the relationship between ESG-related sentiment and the financial performance of Real Estate firms. Using a dataset of ESG-related press releases and news articles from 67 U.S. REITs, we employ three sentiment analysis models—TextBlob, Loughran and McDonald, and FinBERT—to extract sentiment scores at multiple levels. A polarity-based ESG Disclosure Sentiment Score (ESGDSS) is then developed and validated through correlation analysis with MSCI ESG Ratings. Furthermore, a regression-based analysis is conducted to evaluate the impact of ESGDSS on key financial and credit performance indicators. The findings add to our understanding of ESG in the real estate sector and its implications for investors.