摘要
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 purpose of the design of the discretionary accrual models is to reduce the estimation efforts of the non-discretionary part of the accrual to the minimum.This paper reviews the design of the key variables of the model evolution and the selection of the model type,as well as the market application effects of different models.It is found that the capability of the model to separate the accrual has not been improved along with its development;while the practice of the specific accrual that the non-discretionary part is estimated on the basis of the non-accounting factors has less errors than the total accrual that the non-discretionary part is estimated on the basis of the accounting factors.Furthermore study finds that by checking the changes of the model fitting effects after excluding the specific accrual from the aggregate accrual,the manipulation of the excluded specific annual can be verified.Therefore,the exclusive method can be chosen as the breakthrough to lower the estimation error and improve the explanation power of the model.
出处
《当代财经》
CSSCI
北大核心
2012年第12期120-128,134,共10页
Contemporary Finance and Economics
基金
广东省哲学社会科学"十一五"规划项目(GD10YYJ08)
广东省自然科学基金项目(S2012010008876)
关键词
操控性应计
代理变量
估计偏误
discretionary accruals
proxy variables
estimation error