Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Discourse vs emissions: Analysis of corporate narratives, symbolic practices, and mimicry through LLMs

Bertrand Kian Hassani, Yacoub Bahini, Rizwan Mushtaq

TL;DR

This study analyzes corporate climate disclosures for 828 U.S.-listed firms using ClimateBERT-based classifiers to measure narrative maturity across sentiment, commitments, specificity, and net-zero targets. By linking narrative indicators to emissions and firm attributes and applying Gaussian Mixture Model clustering, the paper uncovers a decoupling between tone and targets, and widespread mimetic convergence in disclosures. Larger and higher-emitting firms show more commitments and actions, but these are inconsistently aligned with quantitative targets, suggesting symbolic reporting rather than uniformly credible transition plans. The findings demonstrate the utility of LLMs for ESG narrative analysis and emphasize the need for regulation and assurance to align commitments with verifiable, sector-specific transition strategies.

Abstract

Climate change has increased demands for transparent and comparable corporate climate disclosures, yet imitation and symbolic reporting often undermine their value. This paper develops a multidimensional framework to assess disclosure maturity among 828 U.S.listed firms using large language models (LLMs) fine-tuned for climate communication. Four classifiers-sentiment, commitment, specificity, and target ambition-extract narrative indicators from sustainability and annual reports, which are linked to firm attributes such as emissions, market capitalization, and sector. Analyses reveal three insights: (1) risk-focused narratives often align with explicit commitments, but quantitative targets (e.g., net-zero pledges) remain decoupled from tone; (2) larger and higher-emitting firms disclose more commitments and actions than peers, though inconsistently with quantitative targets; and (3) widespread similarity in disclosure styles suggests mimetic behavior, reducing differentiation and decision usefulness. These results highlight the value of LLMs for ESG narrative analysis and the need for stronger regulation to connect commitments with verifiable transition strategies.

Discourse vs emissions: Analysis of corporate narratives, symbolic practices, and mimicry through LLMs

TL;DR

This study analyzes corporate climate disclosures for 828 U.S.-listed firms using ClimateBERT-based classifiers to measure narrative maturity across sentiment, commitments, specificity, and net-zero targets. By linking narrative indicators to emissions and firm attributes and applying Gaussian Mixture Model clustering, the paper uncovers a decoupling between tone and targets, and widespread mimetic convergence in disclosures. Larger and higher-emitting firms show more commitments and actions, but these are inconsistently aligned with quantitative targets, suggesting symbolic reporting rather than uniformly credible transition plans. The findings demonstrate the utility of LLMs for ESG narrative analysis and emphasize the need for regulation and assurance to align commitments with verifiable, sector-specific transition strategies.

Abstract

Climate change has increased demands for transparent and comparable corporate climate disclosures, yet imitation and symbolic reporting often undermine their value. This paper develops a multidimensional framework to assess disclosure maturity among 828 U.S.listed firms using large language models (LLMs) fine-tuned for climate communication. Four classifiers-sentiment, commitment, specificity, and target ambition-extract narrative indicators from sustainability and annual reports, which are linked to firm attributes such as emissions, market capitalization, and sector. Analyses reveal three insights: (1) risk-focused narratives often align with explicit commitments, but quantitative targets (e.g., net-zero pledges) remain decoupled from tone; (2) larger and higher-emitting firms disclose more commitments and actions than peers, though inconsistently with quantitative targets; and (3) widespread similarity in disclosure styles suggests mimetic behavior, reducing differentiation and decision usefulness. These results highlight the value of LLMs for ESG narrative analysis and the need for stronger regulation to connect commitments with verifiable transition strategies.

Paper Structure

This paper contains 48 sections, 4 figures, 19 tables.

Table of Contents

  1. Introduction
  2. Literature review and hypotheses
  3. Theoretical background
  4. Gap in the literature
  5. Hypothesis
  6. Methodology
  7. Data sources and rationale
  8. Methods
  9. Textual climate disclosure classification
  10. Text extraction and keyword filtering
  11. Clustering and taypology analysis
  12. Results
  13. Interrelations in corporate climate communication
  14. Correlation patterns
  15. Sentiment and commitment:
  16. ...and 33 more sections

Figures (4)

  • Figure 1: Spearman correlation matrix between Climate LLM labels (semantic encoding). Thresholds used: commitment $>$ 40%, specificity $>$ 40%, risk $>$ 30%, opportunity $>$ 30%, net-zero $>$ 30%, reduction $>$ 30%.
  • Figure 2: Elbow Method: Inertia vs Number of Clusters
  • Figure 3: KMeans clusters in PCA space (PC1 vs PC2).
  • Figure 4: KMeans clusters in PCA space (PC1 vs PC3).