Developing a model to predict fraudulent financial reporting

Document Type : Research Paper


Department of Accounting, Neyshabur Branch, Islamic Azad University, Neyshabur, Iran



This paper investigates how well the Beneish and Spathis models can predict fraudulent financial reporting. The coefficients of these two models were adjusted using the logistic regression and the newly adjusted models were investigated for the prediction of fraudulent financial reporting. This research seeks to design a suitable native model to predict possible fraud in financial statements. The statistical population included 99 manufacturing companies listed on the Tehran Stock Exchange (1089 observations) during the years 2009-2019. The results show that the Beneish and Spathis models are not good at predicting fraudulent financial reporting, but their adjusted versions can predict it with an accuracy of 72% and 64%, respectively. The prediction accuracy rate of the extracted model based on the best explanatory variables is 79%, which shows that it is possible to predict and discover fraudulent financial reporting.


Articles in Press, Corrected Proof
Available Online from 10 August 2023
  • Receive Date: 01 August 2022
  • Accept Date: 14 October 2022