论文部分内容阅读
操控性应计模型设计的主旨在于将应计项非操控性部分的估计偏误降至最低。文章回顾了模型演变过程中关键变量的设计与模型形式的选择,以及不同模型的市场应用效果,发现模型分离应计项的能力并没有随着其发展演变而改进,而特定应计基于非会计因素估计非操控性部分的做法,比总应计基于会计因素估计非操控性部分的偏误更低。研究进一步发现,通过检验总体应计排除特定应计后模型拟合效果的改变,可以验证所排除的特定应计项的操控性。因而,排除法形式可以尝试作为降低估计偏误,进而改善模型解释力的突破口。
The main purpose of manipulative accrual model design is to minimize the estimated bias in non-manipulative parts of accruals. The article reviews the design of key variables and the choice of model forms as well as the market application effects of different models. It is found that the ability of the model to separate accruals does not improve as its development evolves, and the specific accruals are based on non-accounting The non-manipulative portion of the factor estimate is less biased than the non-manipulative portion of the total accrual estimate based on accounting considerations. The study further found that handling of the specific accruals that were excluded could be verified by examining the overall accrued exclusion of changes in the fit of specific accrual models. Therefore, the form of exclusion can be used as a breakthrough to reduce the estimation bias and thus improve the explanatory power of the model.