Background This review synthesises research on corporate climate, environmental, ESG, carbon, sustainability, and non-financial disclosure in relation to financing, capital allocation, and investment-related outcomes. It addresses a fragmented evidence base while preserving differences among disclosure constructs, outcomes, and research designs. Methods Using PRISMA 2020-informed procedures, searches were conducted on 3 July 2026 in Scopus and Business Source Complete via EBSCOhost for studies published from 2010 to 2026. Of 2,038 exported-readable records, 66 studies were included after deduplication, language exclusion, screening, retrieval, and full-text assessment. Results The 66-study corpus was retained at PRISMA inclusion and then classified for synthesis: 52 core studies examined focal-firm disclosure, seven provided policy/context evidence, and seven provided mechanism/context evidence. These synthesis roles were analytic classifications, not post-PRISMA exclusions. Financing and cost outcomes were prominent, while the corpus also addressed investment efficiency and real allocation. Findings varied by disclosure construct, dependent-variable orientation, credibility, institutional setting, and research design; they were not interpreted from raw coefficient signs alone. Conclusions The evidence supports a conditional account: specific, credible, and decision-useful disclosure can be associated with financing and allocation outcomes. Causal-effect language is limited to policy or quasi-experimental studies where the reported design, identification assumptions, and supporting diagnostics justify that interpretation.
Systematic Review
[version 1; peer review: awaiting peer review]
https://orcid.org/0009-0003-5260-3356
1, Aldi Jusril Mahendra Pello2, Maria Jessilca J Hera2, [...] Tika Irmawati Ishak2, Nur Afifah Aini2, Atikah Indawati Pua Uda2, Dessy Fitrahtul Hanif2, Sunarti Sunarti2, Sanriana Natali Riwu Djami2, Irwan Josafat Simanjuntak2https://orcid.org/0009-0003-5260-3356
1, Aldi Jusril Mahendra Pello2, [...] Maria Jessilca J Hera2, Tika Irmawati Ishak2, Nur Afifah Aini2, Atikah Indawati Pua Uda2, Dessy Fitrahtul Hanif2, Sunarti Sunarti2, Sanriana Natali Riwu Djami2, Irwan Josafat Simanjuntak21 Department Computer Science and Electronics, Faculty of Mathematics and Natural Sciences, Universitas Gadjah Mada, Yogyakarta, Special Region of Yogyakarta, Indonesia
2 Department of Accounting, Faculty of Economics and Business, Universitas Gadjah Mada, Yogyakarta, Special Region of Yogyakarta, Indonesia
Yayang Matira
Roles: Conceptualization, Data Curation, Formal Analysis, Methodology, Project Administration, Supervision, Visualization, Writing – Original Draft Preparation, Writing – Review & Editing
Aldi Jusril Mahendra Pello
Roles: Data Curation, Investigation, Methodology, Validation, Writing – Review & Editing
Maria Jessilca J Hera
Roles: Data Curation, Investigation, Validation, Writing – Review & Editing
Tika Irmawati Ishak
Roles: Data Curation, Investigation, Validation, Writing – Review & Editing
Nur Afifah Aini
Roles: Data Curation, Investigation, Validation, Writing – Review & Editing
Atikah Indawati Pua Uda
Roles: Data Curation, Investigation, Validation, Writing – Review & Editing
Dessy Fitrahtul Hanif
Roles: Data Curation, Validation, Visualization, Writing – Review & Editing
Sunarti Sunarti
Roles: Investigation, Validation, Writing – Review & Editing
Sanriana Natali Riwu Djami
Roles: Investigation, Validation, Writing – Review & Editing
Irwan Josafat Simanjuntak
Roles: Validation, Writing – Review & Editing
OPEN PEER REVIEW
REVIEWER STATUS AWAITING PEER REVIEW
Corresponding author: Yayang Matira Competing interests: No competing interests were disclosed.
Grant information: The author would like to thank the Indonesia Endowment Fund for Education Agency (LPDP) of the Ministry of Finance of the Republic of Indonesia for the funding support that enabled this research to be carried out and completed properly.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Copyright: © 2026 Matira Y et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. How to cite: Matira Y, Pello AJM, Hera MJJ et al. Climate-related Disclosure, Capital Allocation, and Corporate Investment Efficiency: A Systematic Literature Review and Future Research Agenda [version 1; peer review: awaiting peer review]. F1000Research 2026, 15:1277 (https://doi.org/10.12688/f1000research.187309.1) First published: 04 Aug 2026, 15:1277 (https://doi.org/10.12688/f1000research.187309.1) Latest published: 04 Aug 2026, 15:1277 (https://doi.org/10.12688/f1000research.187309.1)
Climate-related and sustainability disclosure can be economically material to capital providers because it may alter assessments of risk, expected cash flows, financing needs, and the price, availability, capacity, or maturity of external finance. The reviewed studies therefore examine several non-equivalent outcomes, including implied or realised financing prices, financing constraints, debt capacity and maturity, and investment-related allocation. For example, carbon disclosure is associated with a lower implied equity cost in identified subgroups in Xu et al. (2024), while Bonetti et al. (2024) estimate a smaller post-disaster increase in implied cost of capital among prior disclosers. These findings illustrate why disclosure may be relevant to capital-provider decisions without implying that publication automatically produces economic benefits.
The central problem is decision usefulness, not disclosure volume alone. Across the corpus, generic presence or quantity measures, report-content indices, database or rating scores, quantified emissions and targets, and assurance attributes operationalise different constructs ( Table 4). ESG performance is not ESG disclosure; emissions are not disclosure; quantity is not quality; and assurance is one credibility attribute rather than proof that reported content aligns with underlying performance. Evidence can therefore be interpreted comparatively only when topic, reporting vehicle, specificity, credibility, voluntary or mandatory status, and outcome orientation are kept distinct.
The literature is also fragmented across two linked but distinct consequence domains. Financing studies ask whether disclosure is associated with the price, availability, capacity, or maturity of debt or equity finance. Real-allocation studies instead examine investment discipline, portfolio reallocation, or labour allocation. Evidence that disclosure coincides with more favourable financing conditions does not by itself demonstrate that managers select better projects. Across several investment-efficiency studies, residual-based measures proxy deviations from expected investment rather than directly observing project selection (Zadeh et al., 2021; Li et al., 2025; Kouaib, 2022).
This systematic review addresses that fragmentation through a conditional, mechanism-based synthesis rather than a claim of universal disclosure effects. All 66 included studies remain in scope, while their analytic roles are distinguished as 52 core focal-firm-disclosure studies, seven policy/context studies, and seven mechanism/context studies. The synthesis separates formally tested mediators and moderators from inferred or theory-only channels, aligns findings with dependent-variable orientation and design strength, and identifies where financing evidence does and does not extend to real allocation.
• RQ1. How are disclosure constructs conceptualised and measured across topics, reporting vehicles, specificity, and credibility attributes?
• RQ2. What does the evidence show for financing outcomes and real-allocation outcomes?
• RQ3. Which mechanisms and boundary conditions are formally tested, and which channels remain inferred or theory-only?
• RQ4. Which methods and empirical settings dominate the literature, and what gaps follow from them?
This systematic literature review used PRISMA 2020-informed procedures. The methods comprised database searching, eligibility criteria, record and report selection, data extraction, a structured design-feature assessment, descriptive analysis, and research-question synthesis. PRISMA 2020 was used as a reporting framework. The review does not claim independent duplicate screening, consensus adjudication, or inter-rater agreement because these procedures were not established in the available documentation.
The final systematic searches were conducted on 3 July 2026 in Scopus and Business Source Complete via EBSCOhost, with publication-year coverage restricted to 2010–2026. Scopus displayed 1,367 results, of which 1,366 were exported as readable records. Business Source Complete displayed and exported 672 records. In total, 2,038 exported-readable records entered the initial record set.
The database-specific searches combined disclosure terms with financing or investment terms and firm-level terms. The exact database formulations, field restrictions, filters, search dates, and displayed and exported record counts are reported in Table 1. Web of Science Core Collection was considered during protocol development but was not searched because institutional access was unavailable at the time of searching. Publisher platforms were used only for full-text retrieval and were not treated as systematic search databases.
Eligibility criteria addressed publication type, disclosure or policy relevance, financing or investment-outcome relevance, unit of analysis, report retrievability, full-text evidentiary sufficiency, bibliographic identity, and language ( Table 2). No language filter was applied in the EBSCO search (Language: All). After deduplication, 75 non-English records were excluded before title and abstract screening.
These criteria governed study eligibility. The core focal-firm disclosure, policy/context, and mechanism/context labels were assigned only after PRISMA inclusion as analytic classifications for synthesis; they did not constitute additional eligibility criteria or post-inclusion exclusions.
The searches yielded 2,038 exported-readable records. After removal of 537 duplicates, 1,501 records remained. A further 75 non-English records were excluded, leaving 1,426 records for title and abstract screening. Title and abstract prioritisation retained 731 records, of which 218 entered the full-text assessment pool. Of these reports, 177 were retrieved and 41 were not retrieved. Full-text assessment excluded 111 reports, leaving 66 included studies.
In this review, records refer to database entries; reports refer to sought or retrieved full-text documents; and studies refer to the unique included investigations represented by those reports. The available archived screening documentation did not permit retrospective reconstruction of stage-specific exclusion-reason distributions at every transition. No such distribution was inferred.
A standardised extraction workbook captured bibliographic metadata; disclosure construct and measurement; exposure role; outcome and dependent-variable orientation; country or region; sample period; method and identification strategy; theoretical lens; mechanism and moderator status; and synthesis role. Screening summaries and study-level extraction materials were compiled from the available preserved project documentation.
The available documentation does not establish that screening was independently duplicated, resolved through formal consensus, or assessed using inter-rater agreement. Full-text articles were used to verify substantive study characteristics and findings. For traceability, the extraction and supporting materials link reported synthesis statements to the relevant source text, page location, and bounded interpretation. Supporting documentation records 207 claim-level source links across the manuscript synthesis.
Each included study was characterised across ten design-relevant domains: disclosure-construct validity, outcome validity, sample or setting transparency, temporal ordering, confounding controls, endogenous disclosure treatment, identification credibility, robustness or diagnostic reporting, mechanism measurement, and external validity. The assessment was used to discipline interpretation across heterogeneous designs, not to calculate an overall quality score or determine study eligibility.
Each domain was classified as Strong, Moderate, Limited, Not applicable, or Unclear. An estimator label alone, such as fixed effects, matching, GMM, or an instrumental-variable approach, was not sufficient for a Strong classification. Unclear indicates that sufficient information to characterise the domain was not located in the accessible study materials; it does not imply the absence of a safeguard.
Studies were coded by disclosure construct, exposure role, outcome, dependent-variable orientation, method and identification strategy, theoretical lens, mechanism, moderator, and setting. All 66 PRISMA-included studies remained in the review. For synthesis, 52 studies were classified as core focal-firm disclosure studies, seven as policy/context studies, and seven as mechanism/context studies. The policy/context group comprised five mandatory-disclosure policy or shock studies and two broader-policy studies with a disclosure channel. In the mechanism/context group, disclosure was a mediator or secondary variable.
This mutually exclusive 52/7/7 classification was an analytic categorisation for synthesis, not a further eligibility screen or post-inclusion exclusion. Observational estimates were synthesised as associations. Causal-effect language was reserved for studies whose verified design, identification assumptions, and supporting diagnostics justified that interpretation. Direction was harmonised according to dependent-variable meaning; for example, lower financing costs, lower financing constraints, or lower investment inefficiency were interpreted as favourable even where estimated coefficient signs differed.
Preregistration: This systematic review was not preregistered, and no preregistered data-analysis plan was used.
Results first describe the selection of the 66 included studies and their assignment to three analytic synthesis roles. The section then presents descriptive characteristics of the full corpus, followed by the four research-question syntheses. All 66 studies remained part of the included evidence base: 52 were classified as core focal-firm disclosure studies, seven as policy/context studies, and seven as mechanism/context studies. These roles guide interpretation of the evidence; they do not represent additional eligibility criteria or post-inclusion exclusions.
The review included 66 unique studies. For synthesis, these studies were assigned to three mutually exclusive analytic roles: 52 core focal-firm disclosure studies, seven policy/context studies, and seven mechanism/context studies. All 66 remained part of the included evidence base; these roles guided synthesis and did not constitute additional eligibility criteria or post-inclusion exclusions.
Figure 1 presents the study-selection process, while Figure 4 shows how the included studies contribute to the synthesis. Figures 2–5 summarise publication years, geographic settings, synthesis roles, and formal mechanism testing. Their captions identify the relevant denominators and indicate where categories overlap.
Counts report the documented screening pathway ending in 66 unique included studies. Analytic synthesis roles were assigned after PRISMA inclusion and did not constitute additional eligibility criteria.
Study settings are multi-label: a study may be counted for more than one explicitly identified country, while regional and broader geographic settings are reported separately. Counts should not be summed to reconstruct N = 66.
Each of the 66 studies was assigned to one mutually exclusive category: core focal-firm disclosure (n = 52), policy/context (n = 7), or mechanism/context (n = 7).
Of the 66 studies, 43 conducted at least one formal test, whereas 23 conducted none. Among the 43 studies with a formal test, nine tested mediation and 37 tested moderation or other boundary conditions; three studies tested both.
The characteristics tables describe all 66 included studies. The primary synthesis centres on 52 studies in which firm disclosure is the focal exposure. Seven policy/context studies inform interpretation of disclosure-related regulatory, policy, or shock settings, while seven mechanism/context studies inform interpretation of settings in which disclosure is a mediator or secondary variable. These are mutually exclusive analytic synthesis roles, not outcome categories, additional eligibility criteria, or post-inclusion exclusions.
Outcome direction is interpreted according to dependent-variable meaning. Lower financing costs or constraints, greater debt capacity, and lower investment inefficiency are interpreted as favourable despite potentially different raw coefficient signs. Table 3 reports the synthesis roles, Table 4 reports the disclosure and measurement profile, and Table 5 reports the geographic settings of the full corpus.
The evidence shows that ESG performance, ESG disclosure, and climate disclosure are not interchangeable constructs. Across the corpus, disclosure measures vary along four dimensions: topic, reporting vehicle, specificity, and credibility. Topic ranges from broad ESG/CSR and sustainability to environmental, carbon/GHG, and climate-risk information; reporting vehicles include annual, stand-alone sustainability/CSR, integrated, website, database/rating, and mandatory-regime records; specificity ranges from presence indicators and generic scores to quantified emissions, targets, and forward-looking risk; and credibility includes assurance, audit quality, consistency with performance, and greenwashing. For example, a binary carbon-information indicator in annual or ESG reports is not equivalent to a weighted environmental-and-social transparency score assembled from multiple public sources (Xu et al., 2024; Zadeh et al., 2021; Ardina Nuresa and Fatima, 2026; Nosakhare and Monye-Emina, 2026; Harymawan et al., 2022; Ould Daoud Ellili, 2020).
These distinctions matter for decision usefulness. A measure of whether carbon information appears in an annual or ESG report is not equivalent to a weighted environmental-and-social transparency score assembled from multiple public sources (Xu et al., 2024; Zadeh et al., 2021). Similarly, disclosure volume does not necessarily indicate disclosure quality, reported emissions are not equivalent to disclosure itself, and assurance is a credibility attribute rather than proof that reported content fully reflects underlying performance. Findings should therefore not be compared as though they represent one uniform “ESG disclosure” treatment (Garzón Jiménez and Zorio-Grima, 2021; Garzón-Jiménez and Zorio-Grima, 2021; Mathath et al., 2025; Almutairi, 2026; Haninun et al., 2019; Horobet et al., 2024; Qiu et al., 2025b).
Table 4 presents the multi-label taxonomy of topics, reporting vehicles, specificity, and credibility attributes. Table 5 and Figures 2–3 describe the temporal and geographic distribution of the evidence base.
The evidence does not support a universal claim that firms providing disclosure automatically obtain cheaper or easier finance. Instead, parts of the corpus report favourable financing-related associations, including lower costs of capital, reduced financing constraints, and greater debt capacity or longer debt maturity. These associations vary across debt and equity outcomes and according to disclosure credibility, assurance, governance, enforcement, industry exposure, institutional setting, outcome orientation, and research design. Financing price should be distinguished from financing availability, debt capacity, and debt maturity; each should also be distinguished from real capital allocation (García-Sánchez and Noguera-Gámez, 2017; Aleknevičienė and Stralkutė, 2023; Carvalho et al., 2025; Dhar et al., 2026; Hamrouni et al., 2019; He et al., 2013; Kookkaew et al., 2026; Maama and Marimuthu, 2022; Malik and Kashiramka, 2025; Moussa and Elmarzouky, 2024; Nguyen and Duong, 2026; Pirgaip and Rizvić, 2023; Poornima et al., 2025; Rehman et al., 2023; Rizzato et al., 2026; Shad et al., 2020; Truong et al., 2026; Wang et al., 2025b; Wendai et al., 2022; Gerged et al., 2021; Khanchel and Lassoued, 2022).
Verified examples illustrate this distinction. Carbon disclosure is associated with lower cost of capital in some settings, and prior carbon disclosers experienced a smaller post-disaster increase in implied cost of capital in the Fukushima setting (Xu et al., 2024; Bonetti et al., 2024). These findings indicate potentially favourable financing consequences under particular conditions; they do not establish that disclosure uniformly reduces financing costs across firms, markets, or disclosure measures.
The real-allocation evidence concerns investment discipline and portfolio or labour choices rather than financing terms alone. Investment efficiency is commonly operationalised through residual-based measures: some studies use the absolute residual as an inefficiency proxy, whereas others use the residual sign or related classifications to distinguish overinvestment from underinvestment (Zadeh et al., 2021; Li et al., 2025; Kouaib, 2022; Abbas et al., 2025; Anwar and Malik, 2020; Belhoula et al., 2025; da Silva and Vieira, 2025; Duong et al., 2024; Ellili, 2022; Githaiga, 2026; Hai et al., 2022; Hammami and Hendijani Zadeh, 2019; Huang et al., 2023; Jiang et al., 2022; Khalil et al., 2025; Pereira da Silva, 2024; Rustam and Chengxuan, 2022; Zhang et al., 2024). These measures capture deviations from expected investment, not direct observation of project appraisal, project selection, or managerial capital-budgeting quality.
Evidence on mechanisms and allocation should therefore be interpreted cautiously. Zadeh et al. (2021) report partial mediation through forecast error and mediation through forecast dispersion (Zheng and Shen, 2024; Wang et al., 2025a; Zhao et al., 2023). Zhang (2026) reports an approximately 14% reduction relative to the treatment-sample mean, using unaffected insurers as the comparison group. However, direct observation of project-selection mechanisms was uncommon in the coded corpus. Accordingly, these results are retained as associations or design-bounded contrasts rather than as general evidence that disclosure itself causes managers to select superior projects.
The synthesis distinguishes mediation tests, which examine how an association or effect may operate, from moderation and other boundary-condition tests, which examine when, where, or for whom a disclosure–outcome relationship varies. Nine studies formally tested mediation and 37 tested moderation or other boundary conditions; three examined both. Overall, 43 studies conducted at least one formal test, whereas 23 conducted none ( Table 6). The corpus therefore more often addresses when disclosure–outcome relationships vary than how disclosure produces financing or investment consequences through an intervening mechanism.
Formally tested mechanisms include analysts’ information environment, forecast error or dispersion, liquidity, transparency or reputation, and green finance (Zadeh et al., 2021; Xu et al., 2024; Gao et al., 2025; Li et al., 2025; Lu et al., 2025; Meng et al., 2022; Yan et al., 2022; Poornima et al., 2026). Other channels are inferred from study findings or discussed only in theory. They should not be presented as empirically demonstrated mechanisms unless the corresponding indirect path, mediator, or interaction was formally tested.
Boundary conditions reported across the evidence include enforcement, ownership, audit or assurance, environmental exposure, media and analyst capacity, and consistency between disclosure and underlying performance. For example, Fukushima-based estimates show a smaller implied-cost-of-capital increase among prior carbon disclosers. This is a disaster-contingent contrast, however, rather than evidence of a universal disclosure effect (Bonetti et al., 2024; Qiu et al., 2025).
Accordingly, the synthesis uses association language for observational findings and conditional directions. Bounded causal-effect language is reserved for verified mandatory-disclosure, policy-shock, or quasi-experimental designs where the identification assumptions and supporting diagnostics are adequately reported.
Methodological gaps remain substantial. Fixed effects, GMM, matching, lag structures, and instrumental-variable approaches can reduce particular biases, but they do not automatically establish causality or resolve endogenous disclosure choice. The stronger designs in the corpus include the India reporting mandate, China policy reforms, the California insurer database, and the Fukushima shock. Each identifies a setting-specific policy or shock contrast rather than a universal benefit of disclosure (Desai, 2026; Zhou et al., 2026; Zhang, 2026; Bonetti et al., 2024; Akin and Akin, 2026; Ma et al., 2025).
Theoretical and measurement gaps follow from construct heterogeneity. Broad ESG, environmental, carbon, climate, CSR, integrated-reporting, and non-financial measures should not be collapsed into a single disclosure construct, and reporting volume should not be treated as disclosure quality ( Table 4). Table 7 summarises the interpretive research gaps and corresponding future research agenda. Future studies should validate disclosure indices against underlying emissions, targets, and other substantive performance measures; distinguish external assurance from substantive credibility; and test whether apparent disclosure effects reflect information, credibility, monitoring, financing, or managerial-response channels.
Geographic concentration and the use of broad aggregate cross-country samples limit transportability across enforcement regimes, capital-market structures, industries, and institutional settings ( Table 5). Evidence is particularly limited on whether financing-related associations translate into real-allocation outcomes in under-represented settings. Future work should therefore compare enforcement regimes, replicate studies across settings, and examine investment and allocation outcomes separately from financing-price outcomes.
The structured design-feature assessment reinforces these gaps ( Table 8). Construct specification and sample or setting transparency were generally more readily characterised, whereas credible identification and direct mechanism measurement were more limited. Moderate and Limited categories predominated, and Strong was uncommon, appearing only four times in the identification-credibility domain. Unclear indicates that sufficient evidence was not located in the accessible study materials. These categories support domain-specific interpretation; they are not a validated quality instrument, a formal risk-of-bias assessment, or a global ranking of included studies.
Taken together, the most defensible agenda is to use transparent designs with appropriate causal diagnostics, preregister outcome orientation and analytical decisions where feasible, separate disclosure from underlying performance, measure mediators directly, distinguish assurance from credibility, and test whether financing conditions are transmitted into observable allocation decisions.
The review’s theoretical contribution is a mechanism-based and conditional account rather than a claim that disclosure is intrinsically beneficial. Disclosure is more likely to have economic consequences when topic-specific, decision-useful content is sufficiently credible to alter relevant actors’ information sets. Investors, lenders, analysts, boards, and regulators must then process that information through pricing, liquidity, contracting, monitoring, or related responses. These responses may alter financing conditions and, only where they change managerial incentives or resource access, real allocation. This framework distinguishes formally tested mediators and moderators from inferred or theory-only channels and explains why similarly labelled disclosure measures can be associated with different outcomes.
Credibility and measurement validity condition both the information and transmission channels. Reporting volume, broad ESG ratings, topic-specific carbon disclosure, assurance, and consistency between disclosure and underlying performance capture different constructs. Treating them as substitutes can obscure strategic reporting, measurement error, or greenwashing. Firms should therefore prioritise specific, comparable claims linked to measurable performance and targets rather than simply increasing disclosure volume. Investors and lenders should assess construct definition, verification, and consistency with realised performance before incorporating a disclosure measure or score into pricing and stewardship decisions. Auditors and assurers can strengthen process credibility, but assurance should not be interpreted as proof of substantive environmental performance.
Institutional settings further bound transportability. Enforcement, investor protection, ownership structures, state influence, market development, environmental exposure, and analyst or media capacity can affect both disclosure credibility and whether information users act on it. Regulators should therefore pair disclosure requirements with clear definitions, comparability, oversight, and enforcement, while evaluating behavioural and allocation responses rather than assuming that publication alone changes decisions. Geographic concentration and region-level samples also mean that findings should not be transported across jurisdictions without testing institutional differences.
The practical implications are differentiated by actor. Firms need governance and controls that connect disclosures to operational data and capital-planning decisions. Investors and lenders need outcome-oriented due diligence that distinguishes the economic meaning of the dependent variable from the sign of an estimated coefficient. Auditors and assurers should make their scope, criteria, and limitations transparent and assess the reliability of underlying reporting processes. Regulators can improve usability through interoperable definitions and effective enforcement while monitoring strategic compliance and unintended responses. These are conditional recommendations derived from the synthesis, not evidence that every intervention will improve financing or allocation.
Causal interpretation remains constrained because disclosure choice, governance, risk, financing, and investment are often jointly determined. The structured design-feature assessment identifies domain-specific inferential constraints without functioning as a validated quality instrument, formal risk-of-bias tool, overall scoring system, or eligibility criterion. Strong assessments were uncommon, and unresolved assessments were retained as uncertainty rather than treated as absent safeguards. Fixed effects, matching, lag structures, GMM, and instrumental-variable approaches can reduce particular biases but do not automatically eliminate selection, reverse causality, or endogenous disclosure choice.
The review therefore supports a bounded framework: disclosure matters conditionally when it is specific, credible, institutionally supported, processed by relevant actors, and translated through financing or monitoring into allocation. This framework accommodates favourable, null, mixed, and unfavourable findings without treating them as contradictions or pooling incomparable signs. It shifts theory away from a direct disclosure–outcome link and toward explicit tests of information processing, credibility, institutional boundaries, financing transmission, and managerial response.
This review searched Scopus and Business Source Complete through 3 July 2026. Web of Science was not searched because institutional access was unavailable. Although EBSCO was searched with Language set to All, 75 non-English records were excluded after deduplication; coverage is therefore not exhaustive. In addition, the archived screening documentation did not permit retrospective reconstruction of stage-specific exclusion-reason distributions at every transition. The available materials also do not establish independent duplicate screening, consensus adjudication, or inter-rater agreement.
No statistical meta-analysis was undertaken because disclosure measures, estimands, research designs, and dependent-variable orientations were not sufficiently commensurate for defensible coefficient pooling. The structured design-feature assessment is an interpretive aid tailored to this heterogeneous corpus, not a validated measurement instrument or formal risk-of-bias tool. Its categories should therefore not be interpreted as a global ranking of included studies. The synthesis also remains exposed to publication bias in the available literature, construct and measurement disagreement across disclosure indicators, selective or strategic reporting, and endogeneity in many primary studies.
Future reviews should update searches beyond the two available databases where access permits; assess language and search-string sensitivity; validate disclosure indices against underlying performance and targets; report null, mixed, and contradictory findings; distinguish assurance from substantive credibility; and use transparent quasi-experimental diagnostics, direct mediator tests, and cross-jurisdiction replications. These steps can improve future inference, but they cannot retrospectively remove limitations in the reviewed evidence.
This systematic review synthesises 66 studies using a role-sensitive approach: 52 examine focal-firm disclosure as the core exposure, seven provide policy/context evidence, and seven provide mechanism/context evidence. Across disclosure types, favourable, null, mixed, and unfavourable findings vary with the construct measured, outcome orientation, credibility, research design, and institutional setting. Raw coefficient signs therefore do not constitute directly comparable evidence of economic benefit. The synthesis also distinguishes formally tested mediation and moderation from inferred or theory-only channels, while identifying persistent causal, measurement, and transportability constraints.
The review does not support a universal conclusion that disclosure improves financing or allocation outcomes. Specific and credible disclosure can be associated with more favourable financing conditions and, in a narrower set of settings, more disciplined real-allocation outcomes. Publication alone, however, is insufficient, and evidence for a complete disclosure-to-financing-to-allocation pathway remains limited. The most defensible agenda is to align disclosure measures with substantive performance, test credibility and greenwashing directly, identify the actors and mechanisms that process information, strengthen causal designs, and examine whether findings travel across institutional environments. Disclosure is therefore best understood as an input whose economic consequences depend on institutional support, information processing, and transmission through financing or monitoring into allocation.
Ethical approval and consent were not required because this review used bibliographic records and published scholarly literature. It did not involve human participants, personal-data collection, clinical intervention, or unpublished confidential material.
Zenodo: Underlying Data and Reporting Documentation for “Climate-related Disclosure, Capital Allocation, and Corporate Investment Efficiency: A Systematic Literature Review”. https://doi.org/10.5281/zenodo.21474219 [Matira et al., 2026a].
The project contains the following underlying data:
• bibliographic_records_66.csv (bibliographic records for the 66 included studies).
• screening_ledger_included_66.csv (final study-selection record).
• Bibliographic_Records_and_Screening_Ledger_66.xlsx (workbook containing the bibliographic records and final study-selection record).
• Search_Log_Scopus.xlsx and Search_Log_Business_Source_Complete_EBSCOhost.xlsx (database-specific search logs).
• Search_Strategy_Documentation.docx (database search strategies and search scope).
• Figure_1_PRISMA.png (PRISMA study-selection flow diagram).
• README.md and CODEBOOK.md (documentation for the underlying data).
Data are available under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) licence.
Zenodo: Extended Review Data for “Climate-related Disclosure, Capital Allocation, and Corporate Investment Efficiency: A Systematic Literature Review”. https://doi.org/10.5281/zenodo.21475128 [Matira et al., 2026b].
This project contains the following extended data:
• evidence_extraction_matrix_66.csv (study-level evidence-extraction matrix for the 66 included studies).
• rq_synthesis_counts_66.csv (research-question synthesis counts).
• validated_claim_ledger_207.csv (207 short source-linked findings used in the review synthesis).
• Evidence_Extraction_and_Validated_Claim_Ledger.xlsx (workbook containing the extraction and claim-level data).
• PRISMA_2020_Checklist_Completed.docx (completed PRISMA 2020 reporting checklist in editable format).
• PRISMA_2020_Checklist_Completed.pdf (completed PRISMA 2020 reporting checklist in PDF format).
• README.md and CODEBOOK.md (documentation for the extended review data).
Data are available under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) licence.
This review uses PRISMA 2020 principles to report the search strategy, screening flow, eligibility criteria, and transparent synthesis. The PRISMA flow diagram is available in the underlying-data deposit (https://doi.org/10.5281/zenodo.21474219). The completed PRISMA 2020 checklist, study-level extraction framework, and synthesis data are available in the extended-data deposit (https://doi.org/10.5281/zenodo.21475128).
Generative AI disclosure: ChatGPT-5.5 (OpenAI) was used solely for language refinement, formatting support, document organisation, and other non-substantive administrative tasks during manuscript preparation. It was not used to conduct literature searches, screen or select studies, extract or code data, assess study quality or risk of bias, analyse or synthesise evidence, interpret findings, or formulate conclusions. All content was reviewed and approved by the authors, who take full responsibility for the accuracy, integrity, and final content of this manuscript.
The author would like to thank the Indonesia Endowment Fund for Education Agency (LPDP) of the Ministry of Finance of the Republic of Indonesia for the funding support that enabled this research to be carried out and completed properly.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
© 2026 Matira Y et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current Reviewer Status:
AWAITING PEER REVIEW
AWAITING PEER REVIEW
?
Key to Reviewer Statuses VIEW HIDE
ApprovedThe paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approvedFundamental flaws in the paper seriously undermine the findings and conclusions