Detection of Earnings Management. Practical Approach.

Abstract

Many existing earnings management studies establish earnings management detection criteria by examining the statements of every organization under review for a significant period of time. While being suitable for the academia, this approach cannot be adopted by the practitioners, who do not have access to the historical information. The study proposes that the new indicator would be found through examining the existing set of companies and superimposing it to the sets of companies, which should not have the signs of earnings management. Two groups of companies are proposed as candidates for the second samples: companies with high Z-Score and random set of companies. All three samples are examined by using the existing indicators M-Score and fluctuations of operation margin and the conclusion is made that Z-Score is the only indicator, which can easily separate the samples. This study develops a new indicator of presence of earnings management which is based on the assumption that earnings management is the process of fraudulent altering characteristics of the financial statements related to earning. The idea of the study is based on examining the components of DuPont analysis, namely Operating Margin and Asset Turnover. The results show that the sign and the value of the created indicator matches well the definition of earnings management as appearance of change of Operating Margin and Asset Turnover with different signs. The results of the study were tested on few companies involved in the financial scandals of the turn of the century and showed that earnings management could be detected few years prior to when these companies were charged.