Search Results (937)

In corporate green transformation, environmental, social, and governance (ESG) responsibilities in the upstream supply chain are critical components of corporate governance, and disclosure enables outsiders to observe corporate governance practices. This study examines the association between substantive supply chain ESG disclosure (SSD) and green management innovation (GMI), and whether this association varies with three executive characteristics: environmental background, CEO gender, and managerial overconfidence. Using text from the ESG reports of Chinese firms listed on the Shanghai and Shenzhen A-share markets from 2016 to 2024, we employed the large language model DeepSeek to perform sentence-level identification, distinguishing between substantive and rhetorical content in supply chain ESG statements, and aggregated the results to construct a substantive disclosure index. Results show that SSD is positively associated with GMI. This association is stronger in firms with more executives with an environmental background and those led by female CEOs, and weaker in firms with greater managerial overconfidence. This study expands the measurement tools used in disclosure research. The findings also suggest that firms should shift their disclosure focus toward verifiable commitments and provide a new perspective and basis for regulatory evaluation to move from “whether disclosure occurs” to “whether substantive disclosure occurs”. Full article

Classifying corporate environmental, social, and governance (ESG) disclosures is challenging because they combine technical terminology with subtle, domain-specific distinctions. Pretrained Transformers such as ClimateBERT perform well but require many parameters and substantial GPU memory. This paper introduces TC-HGAT, a lightweight hierarchical graph attention network with task-semantic conditioning. TC-HGAT uses frozen FastText-300D embeddings and builds, for each document, a graph of word nodes, virtual cluster nodes, and a task-semantic node obtained by mean-pooling the embeddings of a natural-language task description. On three ESG tasks (Claim, Specificity, and Commitment), TC-HGAT achieves macro-F1 scores of 94.6%, 77.8%, and 79.3%, with the highest mean scores on Claim and Commitment, although no difference from ClimateBERT is statistically significant. Most of the gain comes from the task-semantic node, which improves macro-F1 significantly on all three tasks (by 5.7, 2.4, and 4.5 points), whereas the cluster layer alone gives a modest improvement that is significant only on Claim. Random vectors in the same graph position, including a norm-matched one, do not reproduce this gain, indicating that it depends on the content of the task representation. TC-HGAT trains about 1.4 million parameters, far fewer than BERT-base and ClimateBERT, and needs substantially less peak GPU memory during training. Full article

Major emergencies generate an acute demand for timely and reliable information while simultaneously amplifying the risk of misinformation diffusion on social media platforms. The regulator therefore needs not only to identify individual pieces of misinformation but also to infer a platform’s privately known misinformation-risk type. We develop a mechanism-design model in which verifiable collaborative disclosure and external monitoring jointly reduce unidentified misinformation and support platform risk screening. Three mechanisms are examined: screening based on disclosure–monitoring bundles, disclosure-only regulation, and ex-post auditing supported by credible penalties. Under the specified incentive and feasibility conditions, screening produces different equilibrium regimes as the platform’s commercial return from unidentified misinformation changes. A central result is counterintuitive: the high-risk platform may optimally face lower disclosure and monitoring requirements because this limits the information rent earned by the low-risk platform. We further show that differentiated and uniform disclosure-only arrangements yield the same optimal expected regulatory payoff, while screening can outperform disclosure-only regulation. When screening and ex-post auditing both satisfy their respective optimality and feasibility conditions, their relative performance depends on a critical auditing-cost threshold and distinct participation requirements. These findings provide insights into coordinating information acquisition and risk screening under asymmetric information. Full article

Who performs the work of frontier-AI governance, and what role do commercial specialists play in it? We combine first-party disclosures from 12 developers with at least one publicly announced model whose training compute is estimated to be above ${10}^{25}$ FLOPs with a frozen Crunchbase-derived classifier of commercial AI-governance providers. A systematic verification pass yields 238 realized model- or system-specific governance-provision instances across 10 developers. Collapsing repeated model-level observations produces 99 unique developer–provider/process–function relationships, of which 58 cross an organizational boundary. Evaluation is the most frequently documented function, accounting for 53 of 99 relationships and spanning internal, for-profit/certification, public, nonprofit/research, and unnamed external modes. External evidence remains highly concentrated: OpenAI and Anthropic account for 50 of 58 external relationships. Because the evidence is disclosure-dependent and based primarily on English-language public sources, this concentration should not be read as a population measure of outsourcing. On the commercial side, the frozen classifier identifies 51 canonical specialist firms, 25 of which have positive disclosed funding amounts. These firms have approximately $428.5 million in disclosed historical funding, with 85.3% concentrated in the ten largest. The developer-centered evidence nevertheless contains a broader institutional set of specialist firms, diversified service providers, public institutes, nonprofits, and certification bodies. The results show why venture-oriented provider data and developer disclosures capture different layers of frontier-AI governance and why organizational externality should not be equated with substantive independence or adequate governance capacity. Full article

Under the proprietary cost theory, firms voluntarily disclose information when the benefits of disclosure exceed the costs of revealing sensitive information to competitors or regulators. We hypothesize that management guidance may incur proprietary costs by constraining firms’ ability to pursue aggressive tax strategies and exposing internal performance expectations to tax authorities. Using 18,051 firm-year observations of U.S. public firms from 1993 to 2020, we measure management guidance quality by forecast accuracy and aggressive tax planning by effective tax rates, book–tax differences, discretionary permanent book–tax differences, the ETR differential, and a composite factor score of these measures. We find that management guidance quality is negatively associated with aggressive tax planning. Consistent with the notion that higher-quality management guidance enhances the dissemination of firm-specific information and facilitates regulatory scrutiny, we find that greater IRS attention strengthens the negative association between management forecast accuracy and aggressive tax planning in the subsequent year. To mitigate endogeneity concerns, we exploit an instrumental variable approach with industry peers’ forecast frequency as an instrumental variable and confirm the negative impact of forecast accuracy on tax aggressiveness. Overall, our findings highlight an important role of management guidance quality in corporate tax planning and contribute to the literature on the intersection of corporate disclosures and tax planning strategies. Full article

Corporate carbon comparisons are sensitive to the financial measures, reporting boundaries, and reference firms used. This study examines these choices for 17 automotive groups in a disclosure-based cross-section centered on 2024. A directional distance function combines annual cash investment, a residual operating-cost measure, revenue, and disclosed operational emissions. A secondary model includes reported Scope 3 totals for 16 groups. Holding firm observations fixed, expansion of the operational reference set changes Tesla’s distance from zero to 0.0277 and BYD’s from 0.0127 to 0.0418. Stellantis’ extended distance changes from 0.0860 under strong disposability to zero under activity-specific weak disposability. Calendar-year restrictions, 3720 two-peer deletions, omission of the cost constraint, and observed exchange-rate scenarios further reveal the dependence of the estimates on accounting and model choices. The released calculations were rerun and checked against the archived outputs. The contribution is an empirical assessment of how these choices affect financial–carbon comparisons, with implications for the use of corporate disclosures in screening and oversight. Interpretation remains limited by the small, single-year sample, the accounting relation between cost and revenue, and heterogeneous emissions inventories. A common-category Scope 3 frontier has not been estimated. The distances should therefore guide examination of the underlying comparisons rather than being used as physical efficiency rankings or engineering abatement targets. Full article

Synthetic data generation has become a strategic capability in financial analytics, promising to reconcile the sector’s appetite for data-hungry machine learning with strict privacy, confidentiality, and regulatory constraints. This review provides a structured, evidence-based account of the field that is both conceptual and empirical. On the conceptual side, it explains how the principal generators work, states their objectives formally, and derives the practical consequences of each mechanism for financial data; it then organises evaluation into an operational fidelity–utility–privacy framework, extends it to incorporate the fairness and model-risk requirements imposed by regulated lending and insurance and the validation constraints that differ by data type, and offers explicit guidance on selecting a generator. On the empirical side, it maps the intellectual structure of the field through a science-mapping study of 313 Scopus-indexed documents (1997–2026) retrieved with a four-block Boolean query executed on 27 July 2026 and screened from 591 retrieved records. Performance analysis reveals sustained exponential growth, at an annual rate of 14.1% to 17.5% depending on how the endpoints are set, with 2026 covering only the first 208 days of the year, and 73.8% of output published since 2023. The corpus has accumulated 4359 citations with an h-index of 35 and pronounced inequality (Gini = 0.737; 26.2% of documents uncited), although 96.3% of the uncited documents appeared in 2025 or 2026 and the Gini falls to 0.599 once the analysis is restricted to documents with at least eighteen months of citation exposure, so these figures reflect the citation window as much as scientific impact. Co-word analysis, a strategic thematic map computed from the equivalence index, and an unsupervised clustering pipeline validated using the silhouette, Davies–Bouldin, and Calinski–Harabasz indices, together with seed, subsample, and specification stability tests, identify five overlapping and provisional research fronts: risk modelling and Monte Carlo simulation; fraud and credit-risk detection under class imbalance; privacy-preserving generation; machine-learning analytics and benchmarking; and biomedical simulation methods transferable to finance. The internal validation indices are weak in absolute terms, so these fronts are best read as an exploratory summary of a densely connected intellectual space rather than as sharply delineated lines of enquiry. We conclude that the field’s binding constraint has shifted from generation to certification: realistic financial data can now be produced, but the evidence that would establish a given synthetic dataset as simultaneously realistic, useful and safe—downstream performance on representative financial tasks, quantified disclosure risk, and robustness across market regimes—is still assembled on a study-by-study basis rather than through shared benchmarks. Full article

The weighted average cost of capital (WACC) reflects both the cost and weight of each financing source, so separating them matters for explaining the ESG–firm value relationship. This study examines how ESG disclosure (ESGD) and ESG performance (ESGP) are related to firm value through WACC, with firm size as a moderator. Using 649 firm-year observations from 140 Thai listed firms (2016–2024), we estimate a moderated mediation model with firm-level cluster bootstrapping. ESGD is positively associated with WACC, whereas ESGP is not. Exploratory analysis shows ESGD is positively associated with the equity weight, but not with the cost of equity or cost of debt. Controlling for the equity weight, the ESGD–WACC association becomes insignificant. WACC is positively associated with Tobin’s Q, yet the indirect association of ESGD through WACC is only weakly supported. We find no evidence that this indirect association differs by firm size. With firm fixed effects, ESGD is not significantly associated with WACC, indicating no evidence of a within-firm association. The results suggest, exploratorily, that the ESGD–WACC association may reflect financing composition rather than the cost of each source, underscoring the need to separate these components before interpreting ESG’s role in the cost of capital and firm value. Full article

Environmental, social and governance factors now shape corporate and investment decisions, driven by disclosure regulation and investor demand for comparable non-financial information. Evaluating such performance means combining conflicting, partly qualitative criteria on incommensurable scales, precisely the problem that multicriteria decision analysis was designed to solve. Existing reviews report which methods the field uses, not how these methods are combined, leaving the structural assumptions of the resulting frameworks unexamined. This review addresses that gap. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 statement, 105 peer-reviewed English-language publications from 2016 to 2026 were identified in Scopus (last searched on 30 June 2026). Each study is classified by the role these criteria play in it, and each framework as a three-stage decision system: weighting, standardization onto commensurable scales, and aggregation into a ranking. Design effort is uneven across the three stages: 48 distinct weighting procedures and 54 aggregation rules, against only 12 standardization procedures, two of which cover almost two thirds of the studies. Three quarters of frameworks let performance in one pillar offset performance in another, and almost none report how it would be recalibrated as conditions change. Both findings matter where such frameworks inform investment screening and supplier selection under continually changing disclosure rules. An eight-point research agenda follows. Full article

Corporate supply chain green transformation supports sustainable production and environmental coordination across supply chains. Identifying its key predictors can inform sustainable supply chain management. Using a sample of 1119 Chinese listed firms from 2018 to 2023, this study applies explainable machine learning to examine predictive factors, nonlinear associations, and regional heterogeneity. The results show that firm size, leverage, fixed asset ratio, and operating expense burden make relatively large predictive contributions among internal factors, while industry peer greenness and regional economic development stand out among external factors. The predictive contribution of firm size increases at higher levels, whereas those of the fixed asset ratio and operating expense burden initially rise and then decline. Firm size and industry peer greenness exhibit model-based predictive interactions. Secondary predictors vary across regions: economic development and ownership concentration are prominent in eastern China, trade openness in central China, financing structure and profitability in western China, and ownership type and asset structure in northeastern China. These findings suggest tailoring support to firm resources and regional conditions, strengthening industry collaboration and information sharing, and cross-checking textual disclosures against actual environmental performance. Full article

Environmental disclosure can diverge from rated environmental performance, raising questions about the operating implications of this gap and the role of corporate governance. This study examines these relationships using 32,126 firm-year observations from 4580 Chinese A-share listed firms during 2015–2024. The gap is measured as the difference between industry–year standardized environmental disclosure and Huazheng environmental ratings. Firm and year fixed-effects estimates show a negative average association between the gap and subsequent operating revenue per employee, with sensitivity to historical-performance controls, matching, and outcome measurement. The negative association is stronger among firms with CEO duality: a one-standard-deviation increase in the gap is associated with approximately 2.07% lower revenue per employee in the following year among firms combining the CEO and board-chair positions. The interaction remains negative and significant across the principal revenue specifications and receives weak support at the 10% level under strict accounting value added on a common sample. Component analysis shows that, among firms with CEO duality, disclosure is negatively associated with subsequent revenue per employee conditional on ratings, while ratings are positively associated with the outcome conditional on disclosure. These findings extend organizational decoupling research by identifying leadership structure as a governance condition associated with differences in the operating implications of environmental disclosure–rating divergence. Full article

The growing importance of identifying environmental, social and governance (ESG) risks is driven by increasing regulatory requirements and stakeholder expectations. However, information asymmetry, disclosure–performance gaps and potential greenwashing may limit the reliability of sustainability disclosures. This study examines the potential of financial statements as an additional source for screening ESG risk exposures and proposes an ESG Proxy Mapping Approach based on financial statement indicators. The empirical analysis covers companies listed on the Bulgarian Stock Exchange (68 in total), combining content analysis of publicly available financial and non-financial information with quantitative proxy indicators, comparative analysis and cross-report verification. The findings indicate that the feasibility of identifying potential signals varies across indicators, companies, and sectors. Proxies based on conventional accounting items tend to be more observable than those that require disclosing ESG-specific amounts. However, the mere existence of an observable accounting item or computable ratio does not establish an ESG-related cause. Missing disclosures also constrain the assessment. The findings suggest that, when combined with contextual disclosure analysis, financial-statement-based proxy mapping can support preliminary ESG risk screening, but cannot independently confirm the existence or magnitude of ESG risks. Full article

This article examines the rhetorical and stylistic phenomenon of “highlighting the ending” in prophetic literary units within the Hebrew Bible, situating it within the broader framework of oral–performative culture. Building on the widely accepted view that prophetic texts originated as oral performances, the study emphasizes that written texts represent only partial records of dynamic communicative events involving voice, gesture, and audience interaction. Against this background, the article analyzes a recurrent device in biblical poetry and prose: the marked prominence of the final element of a literary unit. The study surveys scholarship on orality, performance, and textual transmission, demonstrating that meaning in prophetic literature emerges through embodied delivery as much as through linguistic form. It then categorizes a wide range of techniques used to emphasize endings, including expansion or reduction in final lines, syntactic and verbal shifts, chiastic reversals, deviations from repetition patterns, phonetic effects, and climactic disclosure of information. Through numerous examples from prophetic books, the article shows how these techniques function rhetorically to signal closure, enhance memorability, and guide audience interpretation. Ultimately, the prominence of endings is presented as a deliberate and systematic feature rooted in performative communication. Full article

Generative artificial intelligence (GenAI) can expand analytical support but may encourage premature acceptance and insufficient verification. This article presents a multilevel, multimethod dataset generated during a five-week study within ten-week undergraduate courses at a private university in Mexico. The study involved 120 students in 24 pre-existing teams across four sections and two disciplinary contexts. Teams were assigned within sections to two active GenAI conditions: strategic analysis and executive recommendation, or a productive-cognitive-friction condition adding structured uncertainty, simulation, role-based scenarios, stakeholder and ethical analysis, temporal-emotional reflection, and accountability safeguards. Of 118 consenting participants, 114 met the main quantitative-analysis criteria. The release includes de-identified pre/post responses to the 16-item responsible AI-augmented business judgment (RABJ) instrument; neutral individual performance independently assessed by two evaluators; original and neutral team rubric ratings; expert content-validity ratings; numeric qualitative codes; fidelity and assignment metadata; processed data; aggregate outputs; and reproducible notebooks. Technical validation covers integrity, missingness, reliability, interrater and intercoder agreement, content validity, exploratory structure, cluster-aware reproducibility, and disclosure risk. RABJ is a provisional multidimensional framework, not a validated four-factor construct. Direct identifiers, student-generated text, submitted work, and institutional linkage keys are excluded. Full article