Identification of Non-Fragile Variables of Accruals Based on Non-Linear Bayesian Approaches in the Iranian Capital Market
Subject Areas : Financial Econometrics and Quantitative Methods
Azam Rezaian Joybari
1
,
Jomadoordi gorganli davaji
2
,
Majid Ashrafi
3
,
Ali Khamaki
4
1 - PhD student, Department of Accounting , Ali Abad Katoul Branch, Islamic Azad University, Aliabad Katoul, Iran
2 - Assistant Professor, Department of Accounting , Ali Abad Katoul Branch, Islamic Azad University, Aliabad Katoul, Iran
3 - Assistant Professor, Department of Accounting , Ali Abad Katoul Branch, Islamic Azad University, Aliabad Katoul, Iran
4 - Assistant Professor, Department of Accounting , Ali Abad Katoul Branch, Islamic Azad University, Aliabad Katoul, Iran
Keywords: Profit Management, Accruals, Bayesian Averaging,
Abstract :
Purpose: Most accrual models have been developed in the environment of advanced countries (deep capital market) and there is no research that models accruals according to the specific conditions of the country's capital market; As a result, we have tried to identify non-fragile variables affecting accruals in the Iranian capital market by using Bayesian averaging (BMA), dynamic averaging (TVP-DMA) and selective averaging (TVP-DMS) approaches. Research Methodology: The current research method is practical in terms of purpose. The statistical sample of the research includes 171 stock companies in the period of 1390 to 1400. Findings: In this research, 58 variables affecting accruals were included in Bayesian models. The results show that among the BMA, TVP-DMA and TVP-DMS, BVAR and OLS models, the BMA model was determined as the most efficient model. Based on the BMA model, 11 non-fragile variables affecting accruals were identified. Originality/scientific added value: For the first time in internal research, 58 variables affecting accruals have been used to model this index. For comprehensiveness in the modeling of all three types of discretionary accruals models (emphasis on assets and sales, discretionary expenses); Discretionary income model (emphasis on profit, income and cash flow) and hybrid models are used in designing the optimal model.
