重 庆 理 工 大 学

文   献   翻   译

会计收益数据的经验评价
  雷·鲍尔  菲利普·布朗
    对会计实务的有用性进行完全分析方法的局限由会计数据本质上不能被定义的争论来说明。会计数据本质上不能被定义是因为它们缺乏“意义”，从而它们的作用令人质疑。 争论中心在一定程度上源自为适应新经济环境，相应出现会计实务的发展。仅举一些出现问题的领域。随着实务的发展，会计人员需要处理合并、租赁、并购、研发费用、物价波动和税项支出等实务。因为会计缺少一个统一的理论框架，所以在这些会计实务中出现了不一致的现象。其结果是，净收益是不同质部分的累计。因而，净收益被认为是一个“无意义”的数字，跟27张桌子和8把椅子之间的差别没什么不同。在这种观点下，净盈余只能被定义为一系列程序 运用到一系列事件 后得到的结果，没有什么实质内涵。
Canning 观察到
    净盈余的计量结果在任何意义上都不能认为是真实的，除了它是一个数字，是会计人员中止他所采用的程序的应用后得出的结果。
尝试提高计量方式解释能力的分析方法的价值是无争议的。有争议的是这样一个事实：一个分析模型本身没有评估脱离它所隐含计量方式的意义。因此，在没有进行经验检验的基础上，根据会计分析模式得出由于会计收益数据缺乏实质内涵导致它缺乏有用性的结论是不妥的。
    会计收益数据的经验验证需要关于现实世界成果构成的效用实验的协议。因为净收益是一个投资者感兴趣的数字，所以我们用来作为预测标准的结果的是反映在证券价格里的投资决策.净收益数据的内容以及发布时间两项会被用来共同验证会计收益数据的有用性，因为这两项的缺失均会破坏会计收益数据的有用性。
经验检验
    资本理论的最新发展为将证券价格的表现视作会计收益数据有用性的运行测试提供了合理理由。大量令人印象深刻的理论证实，如下的资本市场是有效和无偏的，原因在于如果信息对形成资本资产的价格是有用的，资本市场就会根据这种信息迅速地调整资产的价格，使投资者不能获得更多的非正常报酬。 证据显示，如果证券价格确实根据新信息进行迅速地调整，那么证券价格的变化就会反映信息向资本市场的流动。 可观测的股票价格波动与收益表发布之间的联系可以证明会计收益数据所反映的信息是有用的。
　　　我们采用的将会计收益同股票价格相联系的研究方法就建立在上述理论和通过仅关注影响特定公司股票价格的特定信息得到证据的基础上。 具体来说，我们构建了市场预期收益的两个选择模型来考察当市场预期不准确时市场是如何反应的。
预期盈余变动和未预期盈余变动
　　　根据过去的事实，所有公司的盈余变动具有一致性。一项研究发现，公司平均每股盈余（EPS）变动水平的一半左右与宏观经济效应有关。根据这个证据公司收益从某年到次年的变动中至少有部分可以被预测。在上一年，如果某公司的盈余通过某种特定方式与其他公司的盈余相联系，那么了解了过去的这种特定联系，再加上其他企业当年盈余信息，就可以得到该企业当年盈余的条件期望。因而，除去确定的影响，当前收益所传递的新信息含量通过实际收益变动与条件期望变动的差异得到估计。
但不是所有的这种差异都是新信息。盈余的一些变化源自公司财务和其他政策的改变。我们假定，在第一次估计前，这些变化已经随时间被收益的平均变化所反映。
因为，上述变动的两个组成部分——宏观经济和政策——的影响是同时的，它们的联系可以被联合估计。我们采用的统计流程是：首先采用最小二乘法（OLS）和到上年为止的数据（ ），求出 公司每年的盈余变化（ ）关于市场上其他所有公司（除了 公司）的盈余平均变化（ ）的线性回归系数和截距项（ ）：
　　                （1）
其中，“”表示估计。然后，将第年的市场盈余平均变化代入回归模型，计算出公司在第年的预期盈余变化：
　　                                        
未预期盈余变化，或者说，预测残差（），是盈余变化实际值减去预期盈余变化：
　　                                             （2）
我们假定的当前收益所传递的新信息就是这个预测残差。
市场反应
　　　同样被证明的是，股票价格（由此也有持有股票的报酬率）的变动具有一致性。一项研究估计，在1944年三月到1960年十月期间，股票月报酬率变化的30%-40%左右与市场效应有关。股票报酬中的市场性变化由与所有公司相关的信息的发布所引起。既然我们在评估与个体公司相关的收益表，它的内容和发布时间就应该与剔除了市场效应的公司股票报酬率变化联系起来估计。
　　　市场性信息在投资一美元到公司股票中获得的月回报率上的影响可以由公司普通股月股票价格比关于市场报酬率的线性回归中得出的预测价值来估计：
                               （3）
其中，是公司在第月的月股票价格比，是Fisher的“复合投资业绩指数”的环比[Fisher(1966)]，是公司在第月的股票报酬率的残差。的值是对市场的月报酬率的估计。我们样本中下标m呈现了自1946年1月开始的所有能获取数据的月份数。
　　　等式（3）呈现的普通最小二乘法回归模型（OLS）中的残差计量了基于估计回归参数（）和市场指数的实际报酬率与预期报酬率之间差异的程度。因此，既然发现市场会根据新信息迅速有效地调整，那么这个残差一定代表了公司单独的新信息对持有公司普通股的报酬率的影响。
 

　　　
文献来自，雷·鲍尔和菲利普·布朗，1968，“实证评价会计收益数据”，会计研究期刊，秋天，159-178页.

原文

　　　On accounting practical usefulness fully analysis method limitations by accounting data essentially cannot be defined argument to explain . The essence of accounting data cannot be defined because they lack the " meaning " , and their role in question . Debate centers in part from in order to adapt to the new economic environment , the development of corresponding accounting practice . To name only a few problem areas . With the development of practice , accounting personnel need to deal with , leasing , mergers and acquisitions , R & D , fluctuation of price and tax expenditure . Because of the accounting lacks a unified theoretical framework , so the accounting practice appeared inconsistent phenomenon . As a result, net income is not homogeneous part of the total . Thus , net income is considered to be a " meaningless" digital , with 27 tables and 8 chairs the difference between no different . In this view , net earnings can only be defined as a series of procedures  applied to a series of events  results , no substantial connotation .
　　　
Canning observed
    Net earnings measurement results in any sense are not considered true , except that it is a digital , is accountant the personnel to suspend him the application after the results .
Try to improve the way of measuring ability to explain the analysis method of value is no controversy . Controversial is a such facts : an analytical model itself is not assessed from its implied meaning measurement mode . Therefore , in the absence of experience on the basis of the inspection , according to the accounting analysis model as a result of accounting earnings data lack of substantive connotation causes it to lack of useful conclusion is wrong .
　　　Accounting income data validation experience about real-world results constitute the utility of experimental protocols . Because the net gain is a investors interested in digital , so we used as a predictor of outcome criteria is reflected in the stock price of the investment decision . Net income data content and published two will be used to verify the usefulness of accounting earnings data , because these two are lack of will destroy the data on the usefulness of accounting earnings .
Empirical test
    Capital theory and the latest development of the stock price performance as accounting earnings data useful to run the tests provide reasonable grounds . Many impressive academic proof , the capital market is efficient and unbiased , because if the information on the formation of the capital assets price is useful , the capital market will be based on this information to rapidly adjust the price of assets, so that investors can get more abnormal returns . Evidence shows that , if the prices of securities are based on new information is quickly adjusted , so the price movement of securities will reflect the information to the capital market flows . Observation of the fluctuation of stock price and income statement released the connection between accounting earnings data can prove that reflects the information is useful in .
    We apply the accounting income with stock prices linked to the research method is based on the theory and by only concerned about the impact of specific stock price specific information based on evidence . Specifically , we construct the market expected earnings of two choice model to investigate when the market is expected to less accurate when the market is how to response .
Expected earnings changes and unexpected earnings changes
According to the facts of the past , all of the company 's earnings changes consistent with . One study found that , on average earnings per share ( EPS ) change is about half the level and macroeconomic effect . Based on this evidence , the average earnings per share ( EPS ) change is about half the level and macroeconomic effect . Based on this evidence , the company earnings from a year to change at least portions can be predicted . In the last years , if a company's earnings through a particular manner with other company earnings is linked , so understanding the past the specific connection , plus other enterprises that earnings information , can get the enterprises that the conditional expectation of earnings . Therefore , to remove to determine the effects of current income , the new information content through the actual earnings and conditional expectation change differences are estimated .
　　　But not all of this difference is the new information . Some of the changes from the surplus corporate finance and other policy changes . We assume that , during the first estimates of the former , these changes have been with time was the mean change reflected by income .
　　　Because of these changes, the two components -- macroscopically economy and Policy -- effects are at the same time, they can be combined to estimate. We used statistical process is: first, by using the method of least squares ( OLS ) and to the last date data ( ), and annual earnings change ( ) on the market, all other companies ( in addition to the company 's earnings ) average change () of the linear regression coefficient and the intercept term ( ):
1.
Among them, " " said estimation . Then , the annual market earnings average change into regression model , calculate the company in the years of expected earnings changes :
Earnings change , or, the prediction residuals () , is the actual value of earnings changes expected earnings changes :
2.
We assume that the current yield the new information is the prediction residual .
Market reaction
　　　As proved , the stock price ( which also has to hold the stock return rate ) fluctuant have consistency . One study estimates , in 1944 March to 1960 October , stock month return rate change about 30%-40% and market effect . Stock returns in the market changes from with all company related information released by . Since we evaluate and individual companies related to the income statement , its content and release time should be got rid of and market effect of stock return rate changes linked to estimate :
                           （3）
　　　Market information in one dollar invested into the company stock obtained in the rate of return on common stock of the company can be influenced by the stock price than the market return on rate of linear regression to derive prediction value to estimate : one , is the company in the first month's stock price ratio, is the Fisher " composite performance index " the chain of [Fisher (1966 ) ] , is the company in the first month of the stock return error . The value of market monthly remuneration rate estimation . We sampled a subscript m presented since 1946 January all can obtain data on the number of months .
　　　Equation ( 3) shows the ordinary least squares regression model ( OLS ) in the residual error measurement based on the estimation of the regression parameters ( ) and the market index the actual rate of return and the expected rate of return difference degree . Therefore , since the discovery of the market according to new information quickly and effectively adjusted , so the error must represent the company alone to the new information on the holding of ordinary shares of the company 's remuneration rate influence .

文献来自，雷·鲍尔和菲利普·布朗，1968，“实证评价会计收益数据”，会计研究期刊，秋天，159-178页.Ball, R. and Brown, P., 1968, “An Empirical Evaluation of Accounting Income Numbers”, journal of Accounting Research, Autumn, pp. 159-178