05决策有用的信息观

1、Topic5 The Information Approach to Decision Usefulness1Topic5 The Information Approach to Decision Usefulness1.Overview2.Outline of the Research Problem3.Financial information and Market Response4.The Ball and Brown Study5.Earnings Response Coefficients (ERC)6.A Caveat About the “Best” Accounting Po

2、licy7.The Information Content of Other Financial Statement Information8.Conclusions19-Jul-2221. Overview霍桑实验车间照明实验实验目的：弄清照明强度（自变量）对生产效率（因变量）所产生的影响。实验程序：实验是在被挑选的两组绕线工人中进行的，一组是实验组，一组是控制组；在实验过程中，实验组不断增加照明强度，而控制组照明强度始终保持不变。实验结果：两组的产量均大大增加（前测和后测）了，但增加量几乎相等；无法确定改善照明对生产效率有什么积极影响。Despite the difficulties of

3、 designing experiments to test the implications of decision usefulness, accounting research has established that security market prices do respond to accounting information, that is an examination of empirical research in accounting. 19-Jul-2231.OverviewIf the efficient markets theory and the decisi

4、on theories underlying it are reasonable descriptions to reality on average, we should observe the market values of securities responding in predictable ways to new information.The degree of usefulness for investors can be measured by the extent of volume or price change following release of the inf

5、ormation.The equating of usefulness to information content is called the information approach to decision usefulness of financial reporting, since Ball & Brown (1968).19-Jul-2241.OverviewHowever, what accountants cannot do is claim that the best accounting policy is the one that produces the greates

6、t market response. Why not?The information approach to decision usefulness is an approach to financial reporting that recognizes individuals responsibility for predicting future firm performance and that concentrates on providing useful information for this purpose. The approach assumes securities m

7、arket efficiency, recognizing that the market will react to useful information from any source, including financial statements.19-Jul-2252. Outline of the Research Problem2.1 Reasons for Market Response2.2 Finding the Market Response2.3 Separating Market-Wide and Firm-Specific Factors 2.4 Comparing

8、Returns and Income19-Jul-2262.1 Reasons for Market ResponseConsider the following predictions about investor behavior, in response to financial information:Investors have prior beliefs about a firms future performance, that is, ., which affect the expected return and risk of a firms shares.Upon rele

9、ase of current years net income, certain investors will decide to become more informed, by analyzing the income number.For most of this chapter we will confine financial statement information to reported net income. Why? Is there any other choice?19-Jul-2272.1 Reasons for Market ResponseConsider the

10、 following predictions about investor behavior, in response to financial information:Investors who have revised their beliefs about future firm performance upward will be inclined to buy the firms shares at their current market price, and vice versa.We would expect to observe the volume of shares tr

11、aded to increase when the firm reports its net income (Beaver, 1968). If the investors who interpret reported net income as good news outweigh those who interpret it as bad new, we would expect to observe an increase in the market price of the firms shares, and vice versa.19-Jul-2282.2 Finding the M

12、arket ResponseWhen current years reported net income first became publicly known?Using the date of the firms net income was reported in the financial media such as The Wall Street Journal, and investigating the reactions in a narrow window of a few days surrounding this date.Separating good or bad n

13、ews:The good or bad news in reported net income is usually evaluated relative to what investors expected. This means that researchers must obtain a proxy for what investors expected net income to be.Separating Market-wide and firm-specific factors on share returns:There are always many events taking

14、 place that affect a firms share volume and price. Thus, it is desirable to separate the impacts of market-wide and firm-specific factors on share returns.19-Jul-2292.3 Separating Market-Wide and Firm-Specific FactorThe Market model is widely used to ex post separated market-wide and firm-specific f

15、actors that affect security returns.已实现收益等于期初预期收益j+jRMt加上未期望或异常收益jt。其中：j=(1j)Rf，E(jt)=0，jt0This abnormal return (jt) is also interpreted as the rate of return on firm js shares for time point t after removing the influence of market-wide factors.19-Jul-2210Separating Market Wide and Firm Specific fa

16、ctors If income announcement is good news then we have a positive abnormal share return1119-Jul-22122.3 Separating Market-Wideand Firm-Specific FactorFigure 5.2:Actual return (0.0015) on firm js shares for day 0 (the day of the firms current earnings announcement) is separated into expected return (

17、0.0009) and abnormal return (0.0006). How?Obtain the past Rjt and RMt (proxied, for example, by the Dow Jones Industrial Average index or the S&P/TSX Composite index), and use regression analysis to estimate the coefficients (j and j) of the model. So, we can predict the return on firm js shares wit

18、h j, j and RM0 RM0=(Level of index, end day 0 + Dividends index, day 0) / (Level of index, beginning day 0)1，Sometimes, the dividends are omitted.13Unusual, Non-recurring and Extraordinary ItemsThe extraordinary items must be fully disclosed; otherwise, the market may get an exaggerated impression o

19、f earnings persistence.The last characteristic in the definition was assed in the 1989 revision.Extraordinary items are items that result from transactions or events that have all of the following characteristics: (a) they are not expected to occur frequently over several years; (b) they do not typi

20、fy the normal business activities of the entity; and (c) they do not depend primarily on decision or determinations by management or owners. 19-Jul-22141989 revisionThis revision was designed to resolve the issue of classificatory smoothing, whereby management could smooth (or otherwise manage earni

21、ngs from continuing operations by choosing to classify unusual items above or below the operating earnings line (Barnea, Ronen & Sadan, 1976).However, the nature of the improvement can be questionedTwo related problems arising from this revision:First, if unusual and non-recurring items are not full

22、y disclosed, investors may overestimate the persistence of operating income;Second, and of greater concern, the amounts and timing of the recording of unusual and non-recurring items are subject to strategic manipulation by management.Elliott & Hanna (1996) found a significant decline in the core ea

23、rnings ERC in quarters following the reporting of a large unusual item. Furthermore , the ERC declined further if the firm reported numerous large special items over time. Why?19-Jul-2215Unusual, Non-Recurringand Extraordinary ItemsHierarchy of income numbersNet income before unusual and non-recurri

24、ng items, also called core earnings x xUnusual and non-recurring items x xIncome from continuing operations, also called operating income x xIncome from Discontinuing operations x xNet income x x19-Jul-221619-Jul-2217Other Comprehensive IncomePresented with Income StatementNet income from operations

25、 xxxExtraordinary itemsxxxNet incomexxxOther comprehensive incomexxxComprehensive incomexxxAlternative PresentationAs part of statement of changes in shareholders equityLess transparent, especially if securities markets not fully efficient19-Jul-221819202.4 Comparing Returns and Incomeresearcher can

26、 now compare the abnormal share return (market prices) on day 0 as calculated above with the unexpected component of firms current reported net income (accounting information).If this unexpected net income is good news (that is, a positive one), given securities market efficiency, a positive abnorma

27、l share return constitutes evidence that investors on average are reacting favorably to the expected good news in earnings.To increase the power of the investigation, the researcher may wish to similarly compare a few days on either side of day 0.212.4 Comparing Returns and IncomeIf positive and neg

28、ative abnormal returns surrounding good or bad news are found to hold across a sample of firms, the researcher may conclude that predictions based on the decision theory and efficient securities market theory are supported (that is, accounting information is useful).一种复杂情况是在公司公告盈余时，公司其他特定信息也随之而至简单地把

29、这类公司从样本中剔除出去；另一种复杂情况为了区分市场回报和公司特定回报，对公司的估计用盈余公告后一段期间的数据估计，用其他方法估计和公司特定回报，不区分市场回报和公司特定回报（Easton & Harris, 1991）。并不能保证市场模型充分描述产生股票收益的实际过程Brown & Warner (1980) 得出结论：对于月回报窗口，市场模型比其他可选方法表现得更合理。22（副对角线概率）削弱当期财务报表信息和未来公司业绩之间的关系，称为财务报表中的噪音（noise）或低盈余质量（low earnings quality）。主对角线概率越高，系统越有信息含量（informative），称为

30、透明（transparent）或高质量（high quality）信息系统的信息含量能够被实证检验3.Financial information and Market Response19-Jul-2223Permanent: expected to last indefinitelyTransitory: affecting earrings in the current year onlyPrice Irrelevant: zero persistencyTypes of Earning Events19-Jul-2224Event StudyIt studies the securiti

31、es market reaction to a specific event.事件研究是目前检验半强式有效市场假说的主要方法，用来了解资本市场证券价格与特定事件之间相关性的实证研究若此事件有影响，证券价格波动状况异于无此事件时的表现，产生异常回报应用统计方法检验异常回报状况，以说明此事件是否对证券价格有影响常用事件: 公司盈余公告、新股发行、增发和配股、股票回购或分割、股利分配、兼并收购、盈利预测，以及宏观经济政策变化公告等Ball & Brown (BB, 1968) study 课后自学，下次课提问自行设计一个与BB研究类似 研究构想19-Jul-2225事件及窗口 估计期窗口 事件期窗口

32、样本（分组并归纳样本特征）估计正常和异常收益统计检验窗口长短选择没有固定标准，但数据的可得性会制约窗口长短选择。短窗口从几分钟到几天，长窗口可能涉及几个月到几年短窗口容易避免事件窗内其他事件对证券价格的影响，但短窗口可能错误估计事件窗内预期收益率，而且有些事件的滞后影响可能是短窗口所不能捕获的。因此，近年来长窗口比较流行，但长窗口也存在着诸如遗漏风险因素并错误计量风险、幸存者偏差和数据挖掘偏差等数据问题及统计推断问题正常收益，假设不发生此事件的预期收益，常用计算模型：市场模型、均值调整模型、市场调整模型异常收益，事件期间内证券实际收益与同期正常收益的差19-Jul-2226事件研究法是指运用股

33、票收益率数据来测定某一特定经济事件对公司价值的影响。事件研究法先利用估计期,估计出事件日的期望收益, 由事件期的实际收益扣除期望收益得到非正常收益,再检验样本平均非正常收益是否显著区别于原假设。事件日的期望收益可以由均值调整模型、市场调整模型和市场模型来估计。19-Jul-22274. The Ball and Brown StudyBall & Brown (BB, 1968) began a tradition of empirical markets research in accounting that continues to this day.They were the first

34、 to provide convincing scientific evidence that firms share returns respond to the information content of financial statements.4.1 Methodology and Findings4.2 Causation Versus Association4.3 Outcomes of the BB Study19-Jul-22284.1 Methodology and FindingsBB examined a sample of 261 NYSE firms over ni

35、ne years from 1957 to 1965.BB concentrated on the information content of earnings.BBs first task was to measure the information content of earnings, that is, good news (GN) and bad news (BN). Thus, firms with earnings higher than last years were classified as GN, and vice versa.The next task was to

36、evaluate the market return on the shares of the sample firms near the time of each earnings announcement. This was down according to the abnormal returns procedure illustrated in Figure 5.2. The only difference was BB used monthly returns (daily returns were not available on databases in 1968)BB rep

37、eated their abnormal security market returns calculation for a wide window consisting of each of the 11 months prior to and 6 moths following the month of earnings release (month 0).19-Jul-22294.1 Methodology and FindingsAverage cumulative ones19-Jul-22304.2 Causation Versus AssociationIf a security

38、 market reaction to accounting information is observed during a narrow window of a few days surrounding an earnings announcement, it can be argued that the accounting information is the cause of the market reaction. It cannot be claimed that reported net income caused the abnormal returns during the

39、 11 months leading up to month 0. The most that can be argued is that net income and returns are associated. We will find that the association between share returns and earnings increased as the window widens (Easton, Harris & Ohlson, 1992; Warfield & Wild, 1992)19-Jul-22314.3 Outcomes of the BB Stu

40、dyIt opened up a large number of additional usefulness issues:Whether the magnitude of unexpected earnings is related to the magnitude of the security market response (Beaver, Clarke & Wright, 1979).Since 1968, accounting researchers have studied securities market response to net income on other sto

41、ck exchanges, in other countries, and for quarterly earnings reports, with similar results.The approach has been applied to study market response to the information contained in new accounting standards, auditor changes, etc.Earnings response coefficients (ESC) asks a different question, namely, for

42、 a given amount of unexpected earnings, is the security market response greater for some firms than others?19-Jul-22325.Earnings Response Coefficients(ERC)5.1 Reasons for Differential Market Response5.2 Implications of ERC Research5.3 Measuring Investors Earnings Expectations5.4 SummaryAn earnings r

43、esponse coefficients (ERC) measures the extent of a securitys abnormal market return in response to the unexpected component of reported earnings of the firm issuing that security.19-Jul-22335.1 Reasons for Differential Market ResponseBetaEmpirical evidence of a lower ERC for higher-beta securities

44、was found by Collins & Kothari (1989), and by Easton & Zmijewski (1989).Capital StructureEmpirical evidence of a lower ERC for more highly levered firms was reported by Dhaliwal, Lee & Fargher (1991).Earnings QualityThe higher earnings quality, the higher we would expected the ERC to be.Measurement

45、of earnings quality:Earnings persistenceAccruals qualityOther reasons19-Jul-2234Earnings QualityDescription of event Actual eventWhats meaning of window-dressing?19-Jul-2235Earnings persistenceWe would expect that the ERC will be higher the more the good or bad news in current earnings is expected t

46、o persist in to the future firm performance.Kormedi & Lipe (1987)The measure of persistence was the extent to which earnings changes of the last two years continued into the current year.Ramakrishnan & Thomas (RT, 1991)Different components of net income may have different persistence. This implies t

47、hat accountants should provide lots of classification and detail on the income statement.Permanent, expected to persist indefinitely (ERC1)Transitory, affecting earnings in the current year but not future years (ERC=1)Price irrelevant, persistence of zero (ERC=0) 成功引进新产品，处置产房和设备，资本化开办费，注销研究费19-Jul-2

48、236Higher earnings quality High persistence of earnings and cash flowsHigh predictive ability of earnings and cash flowsHigh earnings response coefficientLow level of earnings managementMore voluntarily disclosureStrong corporate governance19-Jul-2237What should the users be aware of ? Statement use

49、rs must:Understand current financial reporting settings and standards.Differences in accounting methods.Differences in accounting estimates.Differences in standards implementation.Recognize that management may manipulate the financial information.Distinguish between reliable financial statement info

50、rmation and poor quality information.3819-Jul-2238Accruals qualityWe would expect that a higher ERC for higher accruals quality.DeChow & Dichev (2002)Earnings quality depends primarily on the quality of working capital accruals.To the extent current period working capital accruals show up as cash fl

51、ows net period, those accruals are of high quality. A similar argument applies to last periods accruals.Evidence that firms ERCs and share prices respond positively to accrual quality as measured by this procedure is reported by Francis et al (2004, 2005) and Ecker et al (2006).19-Jul-22395.2 Implic

52、ations of ERC ResearchImproved understanding of market response suggests ways that accountants can further improve the decision usefulness of financial statements:Lower informativeness of price for smaller firms implies that expanded disclosure for theses firms would be useful for investors, contrar

53、y to a common argument thatto expand disclosure of the nature and magnitude of financial instruments, including those that are “off-balance-sheet”.the desirability of disclosure of segment information, sinceAlso, MD&A enables the firm to communicate its growth prospect.Disclosure of the components o

54、f net income is useful for investors.19-Jul-22405.3 Measuring Investors Earnings ExpectationsUnder the ideal conditions, expected earnings is simply accretion of discount on opening firm value.When conditions are not ideal:One approach is to project the time series formed by the firms past reported

55、net incomes, that is, to base future expectations on past performance. However, depends on earnings persistence.If earnings are 100% persistent, expected earnings are just last years actual earnings, then unexpected earnings are the change (Ball & Brown, 1968);If earnings are 0 persistent, unexpecte

56、d earnings are equal to the level of current years earnings (Bill Cautions, Example 3.1);Easton & Harris (1991) found both changes in and levels of net income are components of the markets earnings expectation.(To be continued)19-Jul-22415.3 Measuring Investors Earnings ExpectationsUnder the ideal c

57、onditions, expected earnings is simply accretion of discount on opening firm value.When conditions are not ideal:Another source of earnings expectations is analysts forecasts. Since rational investors will presumably use the most accurate forecast.Analysts forecasts are more accurate than time serie

58、s forecasts (Brown et al, 1987; OBrien, 1988);The single most recent earnings forecast provided a more accurate earnings prediction than the average forecast of all analysts following the firm (OBrien, 1988);Analysts forecasts are optimistically, although the bias may hence decreased in recent years

59、 (Kothari, 2001).19-Jul-22425.4 SummaryThe information content of net income can be measured by the extent of security price change or, more specifically, by the size of the securitys abnormal market return, around the time the market learns the current net income.For a given amount of unexpected ne

60、t income, the extent of security price change or abnormal returns depends on factors such asThe empirical results are really quite remarkable.First, they have overcome substantial statistical and experimental design preambles;Second, they supports the theory of securities market efficiency and the d