Abstract:

The paper presents the results of its author's original research into the capability of Beneish M-score to detect accounting fraud using data presented in earnings reports. The research performed is unique in that it uses a sample consisting of enterprises which vary by their legal status, are not listed in regulated markets, and whose managers were known to have been on trial due to suspicion of report falsification. Analysis shows that the Beneish model can detect accounting fraud, although the five-element model gives better results than the eight-element model.