La loi de Benford comme outil d’analyse de détection des anomalies financières

Du 3 juin 2026 au 3 juin 2026

De 13h à 15h, en salle 501 (5^ème étage de la faculté de droit, économie et gestion).
Accessible en distanciel sur l'équipe Teams "Séminaires Granem" (réservé aux membres).

Ce grand séminaire, sur le thème  "La loi de Benford comme outil d’analyse de détection des anomalies financières : fraude fiscale, reporting comptable et enjeux ESG | Benford’s Law as a Tool for Financial Anomaly Detection: Tax Evasion, Accounting Reporting, and ESG Issues" est organisé par l'Axe 3 avec des collègues de l'Université de Macerata, Italie.

• Luisa Scaccia, Raphaella Coppier et Anna-Grazia Quaranta, de l'Université de Macerata (Italie), présenteront deux articles de recherche.

Ce séminaire examine l’usage de la loi de Benford comme outil d’analyse de détection des anomalies financières. Les deux recherches présentées l’appliquent à des contextes complémentaires : d’une part, la construction d’un indicateur de non-conformité fiscale au niveau des entreprises italiennes à partir de données comptables ; d’autre part, l’analyse des liens entre notations ESG et anomalies potentielles dans les états financiers des banques européennes.

En croisant ces deux approches, le séminaire interroge la capacité de méthodes statistiques simples et transparentes à identifier des signaux d’irrégularités financières, tout en discutant leurs apports pour l’audit, la régulation, la fiscalité et la gouvernance des organisations.

Abstracts :

Potentiality of Benford’s law in deriving, a firm-level evasion indicator for Italy
Raffaella Coppier, Elisabetta Michetti & Luisa Scaccia
This paper investigates the potential of Benford’s Law (BL) to detect corporate tax evasion, with a focus on Italy, where tax non-compliance remains a persistent economic challenge. BL predicts the expected distribution of leading digits in naturally occurring numerical data, and significant deviations from this pattern may indicate irregularities, including tax evasion. We apply BL to financial statement data for the years 2014–2022. A firm is classified as potentially non-compliant if its financial data deviate from BL in at least one year. This approach yields a firm-level indicator that can support tax audit targeting and serve as a proxy for tax evasion in empirical research. Such a proxy may help address the scarcity of firm-level data and enable the study of how tax evasion affects firm growth and market distortions. To evaluate its validity, we compare the distribution of potentially non-compliant firms—identified via BL—across regions, sectors, and firm sizes with aggregated data from the Italian Ministry of Economy and Finance and the Revenue Agency, including Synthetic Tax Reliability Indices and other official estimates of tax evasion and the shadow economy. The findings reveal a strong alignment between BL-based results and official indicators, highlighting the potential of BL as a cost-effective, data-driven tool for identifying and studying tax evasion at the firm level.

ESG Ratings and Financial Reporting Anomalies: Evidence from Benford’s Law in the Banking Sector
Authors: Valerio Ficcadenti, Francesca Pampurini, Anna Grazia Quaranta
Abstract. Environmental, Social, and Governance (ESG) ratings are widely used to assess the sustainability and ethical behaviour of financial institutions, yet their ability to capture banks financial reporting’s integrity remains unknown. This doubt may seem unusual since it is well known that ESG ratings do not rely directly on accounting data, but their growing economic, reputational, and regulatory relevance may definitively create indirect incentives for bank managers to adjust financial statements in ways that align reported financial outcomes with declared ESG performance. Such behaviour may particularly affect discretionary accounting areas, including credit risk assessments, loan loss provisions, and regulatory capital determination.
This study investigates the relationship between ESG scores and potential anomalies in banks’ financial statements using a two-step methodological approach. First, a forensic analysis based on Benford’s Law evaluates the overall statistical quality of accounting data. The analysis covers the period 2010–2024 and examines 73 balance sheet and income statement variables for 62 publicly listed European banks. Deviations from expected first-digit distributions are measured using chi-square statistics and Mean Absolute Deviations (MADs). Second, a causal econometric analysis employs multiple fixed effects panel regressions to test the influence of ESG scores on selected discretionary accounting variables (expected credit losses, loan loss provisions and risk weighted assets) during 2020–2024.
The findings reveal, as expected, no significant direct association between ESG ratings and Benford-based anomalies, suggesting that ESG scores do not broadly affect the overall statistical quality of banks’ financial data. However, a significant indirect relationship emerges between ESG scores and risk weighted assets, indicating that ESG pressures may circuitously influence accounting behaviour. The study highlights the need for further research into how institutional and regulatory contexts shape the interaction between sustainability initiatives and financial reporting practices.