Analysis of Accrual Based Models by Sectors in Determining Earnings Management Practices

Abstract

Purpose - Earnings management can be defined as the manipulation of financial statements by managers without breaking the rules. At first glance, it is difficult to detect this manipulation, which is done consciously by managers. Detecting manipulation through “accruals” is one of the frequently used methods. In this study, Jones, Modified Jones and Kothari models were selected among the most important models developed in the previous literature. According to these models, it is aimed to determine which of the models is statistically more significant in determining earnings management practices when sectoral classification is made. Design/methodology/approach - Data were obtained through the COMPUSTAT database. Companies operating in 26 European Union member countries and England, which were active in the database between 1999 and 2020, were scanned and classified as sectoral. GIC-SECTOR Codes classification was used in the sectoral classification. The companies were classified only according to their sector codes, no other classification criteria were used. The data set consists of a total of 28,962 observations. Since the extreme values in the data set are approximately 1% of the total sample, they were not excluded from the sample under the assumption that they would not affect the overall results. Findings - Among the 11 sectors included in the GIC-SECTOR coding, 9 sectors were examined. As a result of the application of the models to the sectors, none of the models within the scope of the study gave statistically significant results in order to determine the earnings management practices of the companies operating in the Health and Communication sectors. In other sectors, at least one of the models yielded significant results. In order for the selected models to work, first of all, the beta coefficients in the models must be calculated. In this study, beta coefficients were calculated separately for each model, and the coefficients of statistically significant models were presented. Discussion - The models used within the scope of the study are the basic models used in determining the application of earnings management through accruals. However, the study shows that there is a gap in the Communication and Health sectors, even if the models are basic models. In other words, there is a need to develop new models for the determination of earnings management practices in the mentioned sectors.