Use of regression models when performing fraud risk assessment procedures in the audit process

Abstract

The article provides an overview of current research in the use of regression models when performing assessment procedures of material misstatement risks due to fraud in the financial statement audit. The authors were reviewing regression models predicting deliberate distortion of financial statements, developed by M. Beneish and J. Jones. In addition, they consider the later modifications to these models, applicable in the course of the audit process to estimate the material misstatement risk on the basis of meso-economic, operational, scaling, and other factors affecting the operations of reporting accountants. The specific features, advantages, disadvantages and the use of different types of regression models in the audit process are described. The criteria for comparison are formulated, and the comparative analysis of adapting the best-known features of regression models to the challenges of fraud risk assessment for financial statements in the audit process is carried out. The conclusions about the possibilities of use and a range of different types of regression models when initiating the fraud risk assessment procedures to the financial statements in the audit process are formulated. The limitations, inherent of such models are explained.