GreenGuard-AU: A Parsimonious RegTech Framework for Greenwashing Detection in Data-Scarce Regulatory Environments

GreenGuard-AU: A Parsimonious RegTech Framework for Greenwashing Detection in Data-Scarce Regulatory Environments | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article GreenGuard-AU: A Parsimonious RegTech Framework for Greenwashing Detection in Data-Scarce Regulatory Environments Pramesh Luitel This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9039579/v2 This work is licensed under a CC BY 4.0 License Status: Posted Version 2 posted You are reading this latest preprint version Show more versions Abstract Corporate greenwashing has grown into a systemic threat to sustainable finance. When firms overstate their environmental credentials, investors get duped and capital flows to the wrong places. Regulators such as the Australian Securi- ties and Investments Commission have responded with enforcement, but the number of confirmed cases remains vanishingly small, which in turn starves supervised machine learning of the labelled data it needs. This paper introduces GreenGuard-AU, an interpretable ensemble designed to screen for greenwashing risk in precisely these data-poor conditions. The framework draws on 38 hand- crafted features spanning textual analysis, longitudinal reporting behaviour, financial alignment, and sector-adjusted risk. We test it on 633 company-year observations drawn from the Australian securities market between 2021 and 2024; only 19 of those carry confirmed enforcement labels, with the rest flagged through ESG rating divergence analysis. The ensemble posts an AUC-ROC of 80.35% ± 1.65%, well ahead of a transformer baseline (63.7% AUC) that buckles under the weight of too few positive examples. A pattern we see again and again in our experiments is what we have come to call the “Over-architecture Penalty”: lean, feature-rich ensembles consistently beat fancier architectures when train- ing labels are this scarce. Perhaps the most striking result is that longitudinal reporting stability, how steady a company’s sustainability language stays from one year to the next, turns out to be a better predictor of regulatory trouble than anything we can extract from a single document. Greenwashing Detection ESG Compliance Ensemble Learning Feature Engineering Natural Language Processing Australian Securities Market Regulatory Technology Interpretable Machine Learning Full Text Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 2 posted You are reading this latest preprint version Show more versions Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9039579","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":603598597,"identity":"021fc6f7-61e6-4525-b1d3-47f731fb52a4","order_by":0,"name":"Pramesh Luitel","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8klEQVRIiWNgGAWjYBACPgbGBgaGAgiHmaECRDI34NXCxsDY2MBgANNyBkQyEtICsgamhbENRBHSIpHc/uCDwZ18/mmHD38unFcbzd8O1PKjYhseLYmNjTMMnlnOuJ2WJj1z2/HcGYcZGxh7ztzGq6WZx+CwAcPtHDNm3m3HchuAWoAuJKDlD1CL/O38z5955xzLnU+UFgagFoPbOQzSvA01uRsIauF52Dizx+CZgeHtNDNpnmMHcjcCtRzE5xd+9vQHH35U3DGQu538+DNPTV3uvPOHDz74UYFbCxQcgDEOo3KJ0VJHhOJRMApGwSgYaQAAzcJdOoJLpPwAAAAASUVORK5CYII=","orcid":"","institution":"Independent Researcher","correspondingAuthor":true,"prefix":"","firstName":"Pramesh","middleName":"","lastName":"Luitel","suffix":""}],"badges":[],"createdAt":"2026-03-05 11:39:25","currentVersionCode":2,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9039579/v2","doiUrl":"https://doi.org/10.21203/rs.3.rs-9039579/v2","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105904812,"identity":"22656f48-6b14-4ef8-901e-f7188f8a0761","added_by":"auto","created_at":"2026-04-01 10:10:37","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2148455,"visible":true,"origin":"","legend":"","description":"","filename":"GreenGuardAu24.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9039579/v2_covered_39acb7f8-2d0e-45f7-af4e-0d327bc96a14.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"GreenGuard-AU: A Parsimonious RegTech Framework for Greenwashing Detection in Data-Scarce Regulatory Environments","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Greenwashing Detection, ESG Compliance, Ensemble Learning, Feature Engineering, Natural Language Processing, Australian Securities Market, Regulatory Technology, Interpretable Machine Learning","lastPublishedDoi":"10.21203/rs.3.rs-9039579/v2","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9039579/v2","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Corporate greenwashing has grown into a systemic threat to sustainable finance.\nWhen firms overstate their environmental credentials, investors get duped and\ncapital flows to the wrong places. Regulators such as the Australian Securi-\nties and Investments Commission have responded with enforcement, but the\nnumber of confirmed cases remains vanishingly small, which in turn starves\nsupervised machine learning of the labelled data it needs. This paper introduces\nGreenGuard-AU, an interpretable ensemble designed to screen for greenwashing\nrisk in precisely these data-poor conditions. The framework draws on 38 hand-\ncrafted features spanning textual analysis, longitudinal reporting behaviour,\nfinancial alignment, and sector-adjusted risk. We test it on 633 company-year\nobservations drawn from the Australian securities market between 2021 and 2024;\nonly 19 of those carry confirmed enforcement labels, with the rest flagged through\nESG rating divergence analysis. The ensemble posts an AUC-ROC of 80.35% ±\n1.65%, well ahead of a transformer baseline (63.7% AUC) that buckles under\nthe weight of too few positive examples. A pattern we see again and again in\nour experiments is what we have come to call the “Over-architecture Penalty”:\nlean, feature-rich ensembles consistently beat fancier architectures when train-\ning labels are this scarce. Perhaps the most striking result is that longitudinal\nreporting stability, how steady a company’s sustainability language stays from\none year to the next, turns out to be a better predictor of regulatory trouble\nthan anything we can extract from a single document.","manuscriptTitle":"GreenGuard-AU: A Parsimonious RegTech Framework for Greenwashing Detection in Data-Scarce Regulatory Environments","msid":"","msnumber":"","nonDraftVersions":[{"code":2,"date":"2026-03-31 18:43:35","doi":"10.21203/rs.3.rs-9039579/v2","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}},{"code":1,"date":"2026-03-10 07:17:25","doi":"10.21203/rs.3.rs-9039579/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"db4147d0-950f-4614-93d2-52eeaf6ff7a2","owner":[],"postedDate":"March 31st, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-10T07:17:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-31 18:43:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v2","identity":"rs-9039579","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9039579","identity":"rs-9039579","version":["v2"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}