[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122453-en":3,"doc-seo-122453-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},122453,1374391974564,"Clementine","https://ap-avatar.wpscdn.com/avatar/14000253aa45c000a9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779874745381141002",8,"Research & Report","Machine Learning Meets Tax Fraud - Insights from Slovakia","Machine learning enables more reliable detection of tax fraud by using a unique dataset of Slovak tax authority audit outcomes with verified cases of tax manipulation. The study addresses a common misclassification challenge and evaluates multiple models to classify tax manipulators using publicly available financial statement indicators. Artificial neural networks, random forests, XGBoost, and support vector machines are compared. Results show XGBoost achieves the best overall F1 score of 0.75 and reaches 0.85 in sector A - Agriculture, supporting audit efficiency improvement.","Machine Learning Meets Tax Fraud: Insights from Slovakia1  \nEduard BAUMÖHL*–Roderik ANTOL**–Tomáš VÝROST*– Tomáš BAČO ***  \nAbstract  \nOne of the most intriguing topics in the field of corporate finance is the detection of tax fraud. We consider a unique dataset of outcomes from Slovak tax authority audits, obtaining valuable insights into verified instances of tax manipulation and avoiding the misclassification problem that is common in this stream of literature. We apply artificial neural networks, random forests, XGBoost, and support vector machines to verify the extent to which we can classify tax manipulators on the basis of publicly available financial statement indicators. Our results show that the XGBoost model demonstrated the highest effectiveness, achieving an F1 score of 0.75 in the full sample, slightly lower scores within the industry groups, and excellent results in sector A – Agriculture, with an F1 score of 0.85. Our results indicate that the use of nowadays commonly known machine learning methods along with standard financial variables can provide a useful tool for tax fraud detection and, as such, can contribute to higher efficiency of tax audits.  \nKeywords: tax frauds, detection models, machine learning, earnings management  \nJEL Classification: C63, G30, G38, K22, K42, M41  \nDOI: [https://doi.org/10.31577/ekoncas.2025.05-06.01](https://doi.org/10.31577/ekoncas.2025.05-06.01)  \nArticle History: Received: February 2025 Accepted: August 2025  \n* Eduard BAUMÖHL, corresponding author – Tomáš VÝROST, University of Economics, Faculty of Commerce, Dolnozemská cesta 1, 852 35 Bratislava, Slovakia; Institute of Economic Research, Slovak Academy of Sciences, Šancová 56, 811 05 Bratislava, Slovakia; Masaryk University, Faculty of Economics and Administration, Department of Finance, Lipová 41a, 602 00 Brno, Czech Republic; e-mail: [eduard.baumohl@euba.sk](eduard.baumohl@euba.sk), ORCID: 0000-0002-5444-7348; [tomas.vyrost@savba.sk](tomas.vyrost@savba.sk), ORCID: 0000-0002-8384-5724  \n** Roderik ANTOL, Comenius University Bratislava, Faculty of Mathematics, Physics and  \nInformatics, Mlynská dolina F1, 842 48 Bratislava, Slovakia; e-mail: [roderik.antol@gmail.com](roderik.antol@gmail.com)  \n*** Tomáš BAČO, Technical University of Košice, Faculty of Economics, B. Němcovej 32, 040 01 Košice, Slovakia; e-mail: [tomas.baco@tuke.sk](tomas.baco@tuke.sk)  \n1 This work was supported by the Slovak Research and Development Agency (grant No. APVV-22- 0126). We are thankful to the Anti-Fraud and Risk Analysis Section of the Financial Directorate of the Slovak Republic for their cooperation. The authors have no competing interests to declare relevant to this article’s content. This work is based on a bachelor’s thesis by Roderik Antol under the supervision of Eduard Baumöhl. All the detailed results and codes are available from the corresponding author upon request.  \nIntroduction  \nIncome taxes are a vital source of revenue for countries worldwide, playing a crucial role in supporting public services and infrastructure. Ensuring the integrity of this income stream is paramount for economic stability and maintaining public trust. However, tax manipulation poses a significant challenge, as it undermines the state’s ability to support its citizens effectively.  \nAs one would expect, companies can try to declare lower revenues or higher costs to reduce their state tax liabilities (Harris et al., 1993) . There are two main strategies for this practice; one is that legal techniques are applied to reduce tax liabilities and maximize after-tax income by exploiting tax loopholes or, generally, conducting strategic planning and structuring of financial activities to exploit tax incentives, deductions, or credits. These legal techniques are usually referred to as earnings management or simply tax avoidance (Beneish, 2001; Ball and Shivakumar, 2008; Huseynov and Klamm, 2012). The other type of strategy is nonlegal, i.e., manipulating financial","cbCaigeC6GZtBQyQ","https://ap.wps.com/l/cbCaigeC6GZtBQyQ","pdf",1926999,1,29,"English","en",105,"# Abstract\n# Introduction\n## Motivation and impact of tax manipulation\n## Tax avoidance vs. tax evasion\n## Existing detection approaches: red flags and logistic regression\n## Machine learning methods for financial manipulation\n## Research gap and model comparison","[{\"question\":\"What dataset and problem does the study use to improve tax fraud detection?\",\"answer\":\"The study uses a unique dataset of verified outcomes from Slovak tax authority audits. It targets a misclassification problem common in the tax-fraud detection literature.\"},{\"question\":\"Which machine learning models are evaluated for classifying tax manipulation?\",\"answer\":\"The study evaluates artificial neural networks, random forests, XGBoost, and support vector machines using publicly available financial statement indicators.\"},{\"question\":\"How effective is the best-performing model and in which sector does it excel?\",\"answer\":\"XGBoost is the most effective model, reaching an F1 score of 0.75 in the full sample. It performs especially well in sector A - Agriculture with an F1 score of 0.85.\"}]","Machine Learning Meets Tax Fraud - Insights from Slovakia | PDF",1785810719,73,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"machine-learning-meets-tax-fraud-insights-from-slovakia","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/machine-learning-meets-tax-fraud-insights-from-slovakia/122453/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What dataset and problem does the study use to improve tax fraud detection?","Question",{"text":75,"@type":76},"The study uses a unique dataset of verified outcomes from Slovak tax authority audits. It targets a misclassification problem common in the tax-fraud detection literature.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which machine learning models are evaluated for classifying tax manipulation?",{"text":80,"@type":76},"The study evaluates artificial neural networks, random forests, XGBoost, and support vector machines using publicly available financial statement indicators.",{"name":82,"@type":73,"acceptedAnswer":83},"How effective is the best-performing model and in which sector does it excel?",{"text":84,"@type":76},"XGBoost is the most effective model, reaching an F1 score of 0.75 in the full sample. It performs especially well in sector A - Agriculture with an F1 score of 0.85.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]