Unaudited ESG reporting can inflate apparent firm value while hiding earnings management, creating an 'ESG paradox' that corrodes trust. Standardized disclosure and audit/assurance are urgently needed — both to protect investors and to prevent AI models from learning on noisy or manipulated ESG signals.

Main Finding

ESG has become central to financial-accounting decision-making—affecting financial performance, firm value, and cost of capital—but it embodies an “ESG paradox”: while ESG disclosure can signal long‑term value and legitimacy, it also creates opportunities for earnings management, creative accounting, and greenwashing. Strengthened internal controls, auditing, ESG ratings, and regulatory intervention are urgently needed to preserve the credibility of ESG information.

Key Points

• Scope and focus
  • The essay reviews literature on sustainability in financial accounting, concentrating on ESG (with CSR treated as an ESG proxy).
  • Dominant research questions examine how ESG affects financial performance, firm value, and cost of capital, and how accounting practices (e.g., earnings management) interact with ESG reporting.
• Positive and negative roles of ESG
  • ESG investment and disclosure can support long‑term performance, legitimacy, and investor decision‑making.
  • ESG is costly and may reduce short‑term profits, creating tension between management and short‑term investors.
• ESG as a signaling and governance instrument
  • Firms use ESG disclosure to signal stability and commitment to stakeholders; ESG ratings (e.g., via exchanges partnering with providers like Sustainalytics) are becoming important monitoring tools.
• Risks and opportunism
  • ESG reporting can be manipulated to mask poor financial performance or to cover earnings management and other forms of creative accounting.
  • Real earnings management (REM) can co‑occur with sustainability reporting, undermining transparency and stakeholder trust.
• Governance, controls, and regulation
  • Effective internal control systems and rigorous external audit are highlighted as critical to ensuring reliability of ESG data.
  • The paper stresses the need for regulatory intervention and improved standards for ESG reporting and ratings.
• Gaps identified
  • The reviewed literature often emphasizes investor perspectives and financial outcomes; it under‑explores behavioral factors, fraud dynamics, and the micro‑mechanics of how accounting opportunism and ESG reporting interact.

Data & Methods

• Methodology type: qualitative narrative literature review (essay style).
• Data sources: peer‑reviewed journal articles in financial accounting, sustainable finance, and corporate governance (searches likely in databases such as Scopus and ScienceDirect).
• Search/selection approach: keyword searches (e.g., “ESG reporting,” “sustainability accounting,” “earnings management”), screening by relevance, timeframe (recent decade emphasis), and journal reputation.
• Analytical procedure:
  • Extraction and categorization of findings into themes.
  • Two primary themes emerged: (1) ESG implications for financial performance/firm value/cost of capital, and (2) relationship between ESG reporting and questionable accounting practices (earnings management).
  • Synthesis focused on tensions between long‑term ESG benefits and risks of opportunistic behavior.
• Limitations acknowledged by the authors:
  • No quantitative meta‑analysis; narrative review cannot estimate effect sizes.
  • Limited deep coverage of behavioral mechanisms, fraud case studies, and firm‑level empirical tests in the review.

Implications for AI Economics

• AI as a tool to improve ESG data quality and detection of opportunism
  • Machine learning (ML) and natural language processing (NLP) can extract and standardize ESG disclosures from heterogeneous reports, increasing comparability and reducing manual reporting costs.
  • Anomaly detection and forensic ML can flag patterns consistent with earnings management or greenwashing (e.g., divergence between financials and ESG narratives; sudden shifts in disclosure language).
• Market and valuation implications
  • More reliable, AI‑enhanced ESG signals could be incorporated into asset‑pricing models and cost‑of‑capital estimates, affecting portfolio allocation and risk premia for sustainability factors.
  • Conversely, AI‑derived ESG scores may become new targets for strategic manipulation—creating second‑order strategic behavior where firms learn to game models.
• Policy, regulation, and algorithmic governance
  • Regulators should require transparency and auditability for AI systems used in ESG scoring and surveillance (model provenance, data sources, feature importance, robustness checks).
  • Standards for ESG data labeling and model validation would reduce model arbitrage and limit perverse incentives.
• Research opportunities in AI economics
  • Causal ML methods to estimate the true causal effect of ESG actions on firm performance and cost of capital (addressing selection and endogeneity).
  • Game‑theoretic and mechanism‑design analyses of how firms respond to algorithmic ESG ratings and automated surveillance.
  • Behavioral and incentive models of manager responses to AI‑based detection tools and public ESG metrics.
• Practical recommendations
  • Deploy hybrid systems: combine AI screening tools with human audit to balance scalability and interpretability.
  • Use explainable AI (XAI) methods to produce audit‑friendly ESG metrics that regulators, auditors, and stakeholders can verify.
  • Encourage public datasets and open benchmarks for ESG scoring models to improve robustness and reduce divergence across providers.

Short takeaway: the paper’s call for stronger controls and regulation of ESG information maps directly onto urgent needs in AI economics—better AI tools can raise ESG data quality and monitoring capacity, but they must be designed, validated, and governed to avoid creating new avenues for strategic manipulation.