基于成本粘性的盈利预测及其精度检验  

The Earnings Forecasting Model Based on Cost Stickiness and It's Forecast Accuracy Test

作　　者：苏文兵[1] 李心合[1] 段治翔[2] 

机构地区：[1]南京大学会计与财务研究院,江苏南京210093 [2]国家外汇管理局中央外汇业务中心,北京100032

出　　处：《数理统计与管理》2012年第5期930-940,共11页Journal of Applied Statistics and Management

基　　金：国家自然科学基金项目(70572049)

摘　　要：本文以2007-2009年沪深2652家A股上市公司为样本,验证了成本粘性的存在性,并以2008年的相关财务数据为基础构建盈利预测模型,对2009年的公司盈利进行预测,通过与当年各公司的实际利润相比较,发现基于成本粘性的盈利预测模型具有更高的预测能力,其预测精度显著高于其他常用预测模型,表明将诸如成本习性等管理会计的工具引入财务会计研究领域具有重大意义,可以提高会计信息的决策相关性。The paper starts from testifying the existence of cost stickiness using 2652 observations of listed companies＇ data in 2007-2009. Then we estimate the parameters of all our four forecast models using samples in the period in 2008, through which we obtain ROE forecasts for 2009. By comparing the actual and forecasted ROE, we find that our CS model provides substantial improvement in forecast accuracy over the other three models that use only line items in the financial statements. The substantial forecast accuracy improvement CS model also points to the rich potential of decision-making relevance increasing for future research integrating heretofore disparate streams of work in managerial and financial accounting.

关 键 词：成本粘性 盈利预测 预测精度 

分 类 号：O212[理学—概率论与数理统计]