Prediction of earnings quality using artificial neural network: evidence from Tehran stock exchange (Case study: companies admitted to the Tehran Stock Exchange)

Subject Areas : Financial Economics

Loghman Hatami Shirkouhi ¹ , Soghra Barari Nokashti ² , Maryam Ooshaksarae ³

1 - Assistant Professor, Department of Accounting, Rasht Branch, Islamic Azad University, Rasht, Iran
2 - Assistant Professor, Department of Accounting, Rasht Branch, Islamic Azad University, Rasht, Iran
3 - Assistant Professor, Department of Accounting, Rasht Branch, Islamic Azad University, Rasht, Iran

Received: 2023-03-12 Accepted : 2024-10-07

Keywords: earnings quality, neural network, accrual quality,

Abstract :

This study tends to investigate the effective factors on earnings quality and to help decision makers, including investors, to identify the quality of the earnings reported by management in financial statements based on artificial neural network (ANN). The method used in this study is multilayer perceptron ANN. The time scope of the study begins from 2010 to 2022; the statistical population of the study includes all firms listed in Tehran Stock Exchange. In order to identify and predict earnings quality, parameters related to corporate governance, dividend policy, debt financing, conservatism, and other effective factors were studied. Based on the results, the network had the best performance in earnings quality prediction with a precision of 93.96%. The results also show that dividend policy, board independence, audit committee independence, conservatism, organizational ownership, and debt financing have the greatest effect on earnings quality, respectively. The results show that dividend policy (25%), board independence (21%), and audit committee independence (15%), respectively, are the most important variables in earnings quality prediction. The results of the model can be useful for earnings quality prediction by investors, shareholders, creditors, and other interested groups.