NLP-Based Quantification of ESG in Sustainability Reports and Firm-Specific Risk: Evidence from Borsa İstanbul

Yunus Emre Akdoğan and Bayram Aydın
Additional contact information
Yunus Emre Akdoğan: Yozgat Bozok University
Bayram Aydın: Yozgat Bozok University

Business and Economics Research Journal, 2025, vol. 16, issue 4, 417-433

Abstract: This study examines the effect of the sentiment used in ESG (environmental, social, and governance) communication in corporate sustainability reports on firm-specific (idiosyncratic) risks in the Turkish stock market. The analysis focuses on 28 firms listed on the Borsa Istanbul Sustainability Index (XUSRD) between 2014 and 2023, based on a panel of 280 firm-year observations drawn from their publicly available 10-year sustainability reports. First, ESG disclosures in corporate sustainability reports were classified using natural language processing (NLP) techniques and transfer learning. Sentiment analysis was performed for each ESG dimension, and sentiment indices were created based on the analysis results. The data obtained were then analyzed using panel ARDL and panel causality test to test the effect of ESG sentiment on firm risks. The findings reveal that among the ESG dimensions, environmental and governance components play a particularly important role in reducing firm-specific idiosyncratic risk. Also, the results demonstrate the usability of AI-supported analyses in investment strategies and the economic benefit potential of ESG-focused corporate communication. In this context, ESG sentiment is critical not only from a social responsibility perspective but also in terms of risk management and investment decisions.

Keywords: Sustainability Reporting; ESG Sentiment; Idiosyncratic Volatility; Natural Language Processing; BERT-ESG; Economic Growth (search for similar items in EconPapers)
JEL-codes: C23 C80 E44 G10 G32 (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.berjournal.com/nlp-based-quantificatio ... -from-borsa-istanbul

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:ris:buecrj:021717

DOI: 10.20409/berj.2025.475

Access Statistics for this article

Business and Economics Research Journal is currently edited by Adem Anbar

More articles in Business and Economics Research Journal from Bursa Uludag University, Faculty of Economics and Administrative Sciences Contact information at EDIRC.
Bibliographic data for series maintained by Adem Anbar ().