چکیده:
The specific nature of structural relationships compared to operational relationships in accounting has posed questions regarding the efficiency and economic feasibility of panel analysis versus pooled analysis in models based on structural and operational relationships for researchers. In this article, by using the literature of the functional stability hypothesis and testing said hypothesis, an attempt is made to answer questions regarding the efficiency and economic feasibility of panel analysis. The identification method used by the researchers involves implementing 9 regression models based on panel analysis and one regression model based on pooled analysis regarding the information of 153 companies over 7 years (1382-1389), totaling 1071 company-years. The reason for using the functional stability hypothesis is that the methodology for testing this hypothesis includes both types of models—profit component persistence models (structural relationship) and return prediction models (operational relationship). The research results indicate no discrepancy between the results of panel and pooled analyses regarding the profit component persistence model, while differences exist between the results of panel and pooled analyses regarding the return prediction model.
خلاصه ماشینی:
According to the theoretical framework presented in this article, the functions of panel data analysis in accounting for the effects of other influencing variables, such as year and company, are weaker in models based on structural relationships than in models based on practical relationships; consequently, the efficiency and economic cost-effectiveness of panel data analysis, given the costs and disadvantages that this type of analysis entails, becomes ambiguous compared to pooled analyses.
In the second section, explanations are provided regarding the functional stability hypothesis and the methodology used for this hypothesis in the research literature, and we state the reason for using this test in analyzing the efficiency and economic cost-effectiveness of panel analysis.
Therefore, given the theoretical framework presented, researchers hypothesize that in models based on structural relationships, compared to models based on practical relationships, the effects of year and company are less important, and the results of panel and pooled analyses do not have a significant difference from each other.
Statistical findings of the research Earnings components persistence model - structural relationship-based model Table number three shows the model coefficients, t-statistic, F-test statistic, adjusted R2, and finally the Durbin-Watson statistic for 9 different cases of panel analysis obtained under the assumption that slope coefficients are fixed, and one case of pooled analysis.
Results of panel and pooled analysis of the earnings persistence model based on structural relationships Cross-section Period Coefficient t-statistic F-test Durbin Row (Company) (Year) Intercept Accruals Cash Flows Intercept Accruals Cash Flows Statistic Adjusted R2 Watson 1 None None 0.
(
, 
, 
)
