決策有用的信息觀

Topic5TheInformationApproachtoDecisionUsefulnessTopic5TheInformationApproachtoDecisionUsefulness1.Overview2.OutlineoftheResearchProblem3.FinancialinformationandMarketResponse4.TheBallandBrownStudy5.EarningsResponseCoefficients(ERC)6.ACaveatAboutthe“Best”AccountingPolicy7.TheInformationContentofOtherFinancialStatementInformation8.Conclusions19-Jan-2321.Overview霍桑實驗——車間照明實驗實驗目的：弄清照明強度（自變量）對生產(chǎn)效率（因變量）所產(chǎn)生的影響。實驗程序：實驗是在被挑選的兩組繞線工人中進行的，一組是實驗組，一組是控制組；在實驗過程中，實驗組不斷增加照明強度，而控制組照明強度始終保持不變。實驗結(jié)果：兩組的產(chǎn)量均大大增加（前測和后測）了，但增加量幾乎相等；無法確定改善照明對生產(chǎn)效率有什么積極影響。Despitethedifficultiesofdesigningexperimentstotesttheimplicationsofdecisionusefulness,accountingresearchhasestablishedthatsecuritymarketpricesdorespondtoaccountinginformation,thatisanexaminationofempiricalresearchinaccounting.19-Jan-2331.OverviewIftheefficientmarketstheoryandthedecisiontheoriesunderlyingitarereasonabledescriptionstorealityonaverage,weshouldobservethemarketvaluesofsecuritiesrespondinginpredictablewaystonewinformation.Thedegreeofusefulnessforinvestorscanbemeasuredbytheextentofvolumeorpricechangefollowingreleaseoftheinformation.Theequatingofusefulnesstoinformationcontentiscalledtheinformationapproachtodecisionusefulnessoffinancialreporting,sinceBall&Brown(1968).19-Jan-2341.OverviewHowever,whataccountantscannotdoisclaimthatthebestaccountingpolicyistheonethatproducesthegreatestmarketresponse.Whynot?Theinformationapproachtodecisionusefulnessisanapproachtofinancialreportingthatrecognizesindividualsresponsibilityforpredictingfuturefirmperformanceandthatconcentratesonprovidingusefulinformationforthispurpose.Theapproachassumessecuritiesmarketefficiency,recognizingthatthemarketwillreacttousefulinformationfromanysource,includingfinancialstatements.19-Jan-2352.OutlineoftheResearchProblem2.1ReasonsforMarketResponse2.2FindingtheMarketResponse2.3SeparatingMarket-WideandFirm-SpecificFactors2.4ComparingReturnsandIncome19-Jan-2362.1ReasonsforMarketResponseConsiderthefollowingpredictionsaboutinvestorbehavior,inresponsetofinancialinformation:Investorshavepriorbeliefsaboutafirm’sfutureperformance,thatis,…….,whichaffecttheexpectedreturnandriskofafirm’sshares.Uponreleaseofcurrentyear’snetincome,certaininvestorswilldecidetobecomemoreinformed,byanalyzingtheincomenumber.Formostofthischapterwewillconfinefinancialstatementinformationtoreportednetincome.Why?Isthereanyotherchoice?19-Jan-2372.1ReasonsforMarketResponseConsiderthefollowingpredictionsaboutinvestorbehavior,inresponsetofinancialinformation:Investorswhohaverevisedtheirbeliefsaboutfuturefirmperformanceupwardwillbeinclinedtobuythefirm’ssharesattheircurrentmarketprice,andviceversa.Wewouldexpecttoobservethevolumeofsharestradedtoincreasewhenthefirmreportsitsnetincome(Beaver,1968).Iftheinvestorswhointerpretreportednetincomeasgoodnewsoutweighthosewhointerpretitasbadnew,wewouldexpecttoobserveanincreaseinthemarketpriceofthefirm’sshares,andviceversa.19-Jan-2382.2FindingtheMarketResponseWhencurrentyear’sreportednetincomefirstbecamepubliclyknown?Usingthedateof

thefirm’snetincomewasreportedinthefinancialmediasuchasTheWallStreetJournal,andinvestigatingthereactionsinanarrowwindowofafewdayssurroundingthisdate.Separatinggoodorbadnews:Thegoodorbadnewsinreportednetincomeisusuallyevaluatedrelativetowhatinvestorsexpected.Thismeansthatresearchersmustobtainaproxyforwhatinvestorsexpectednetincometobe.SeparatingMarket-wideandfirm-specificfactorsonsharereturns:Therearealwaysmanyeventstakingplacethataffectafirm’ssharevolumeandprice.Thus,itisdesirabletoseparatetheimpactsofmarket-wideandfirm-specificfactorsonsharereturns.19-Jan-2392.3SeparatingMarket-Wide

andFirm-SpecificFactorTheMarketmodeliswidelyusedtoexpostseparatedmarket-wideandfirm-specificfactorsthataffectsecurityreturns.已實現(xiàn)收益等于期初預期收益αj+βjRMt加上未期望或異常收益εjt。其中：αj=(1－βj)Rf，E(εjt)=0，εjt≠0Thisabnormalreturn(εjt)isalsointerpretedastherateofreturnonfirmj’ssharesfortimepointtafterremovingtheinfluenceofmarket-widefactors.19-Jan-2310SeparatingMarketWideandFirmSpecificfactorsIfincomeannouncementisgoodnewsthenwehaveapositiveabnormalsharereturn01-Jan-23122.3SeparatingMarket-Wide

andFirm-SpecificFactorFigure5.2:Actualreturn(0.0015)onfirmj’ssharesforday0(thedayofthefirm’’scurrentearningsannouncement)isseparatedintoexpectedreturn(0.0009)andabnormalreturn(0.0006).How?ObtainthepastRjtandRMt(proxied,forexample,bytheDowJonesIndustrialAverageindexortheS&P/TSXCompositeindex),anduseregressionanalysistoestimatethecoefficients(αjandβj)ofthemodel.So,wecanpredictthereturnonfirmj’’sshareswithαj,βjandRM0RM0=(Levelofindex,endday0+Dividendsindex,day0)/(Levelofindex,beginningday0)－1，Sometimes,thedividendsareomitted.Unusual,Non-recurringandExtraordinaryItemsTheextraordinaryitemsmustbefullydisclosed;otherwise,themarketmaygetanexaggeratedimpressionofearningspersistence.Thelastcharacteristicinthedefinitionwasassedinthe1989revision.Extraordinaryitemsareitemsthatresultfromtransactionsoreventsthathaveallofthefollowingcharacteristics:(a)theyarenotexpectedtooccurfrequentlyoverseveralyears;(b)theydonottypifythenormalbusinessactivitiesoftheentity;and(c)theydonotdependprimarilyondecisionordeterminationsbymanagementorowners.01-Jan-23141989revisionThisrevisionwasdesignedtoresolvetheissueofclassificatorysmoothing,wherebymanagementcouldsmooth(orotherwisemanageearningsfromcontinuingoperationsbychoosingtoclassifyunusualitemsaboveorbelowtheoperatingearningsline(Barnea,Ronen&Sadan,1976).However,thenatureoftheimprovementcanbequestioned……Tworelatedproblemsarisingfromthisrevision:First,ifunusualandnon-recurringitemsarenotfullydisclosed,investorsmayoverestimatethepersistenceofoperatingincome;Second,andofgreaterconcern,theamountsandtimingoftherecordingofunusualandnon-recurringitemsaresubjecttostrategicmanipulationbymanagement.Elliott&Hanna(1996)foundasignificantdeclineinthecoreearningsERCinquartersfollowingthereportingofalargeunusualitem.Furthermore,theERCdeclinedfurtherifthefirmreportednumerouslargespecialitemsovertime.Why?01-Jan-231501-Jan-2316Unusual,Non-Recurring

andExtraordinaryItemsHierarchyofincomenumbersNetincomebeforeunusualandnon-recurringitems,alsocalledcoreearningsxxUnusualandnon-recurringitemsxxIncomefromcontinuingoperations,alsocalledoperatingincomexxIncomefromDiscontinuingoperationsxxNetincomexx01-Jan-231701-Jan-2318OtherComprehensiveIncomePresentedwithIncomeStatementNetincomefromoperationsxxxExtraordinaryitemsxxxNetincomexxxOthercomprehensiveincomexxxComprehensiveincomexxxAlternativePresentationAspartofstatementofchangesinshareholders’equityLesstransparent,especiallyifsecuritiesmarketsnotfullyefficient19202.4ComparingReturnsandIncomeresearchercannowcomparetheabnormalsharereturn(marketprices)onday0ascalculatedabovewiththeunexpectedcomponentoffirm’’scurrentreportednetincome(accountinginformation).Ifthisunexpectednetincomeisgoodnews(thatis,apositiveone),givensecuritiesmarketefficiency,apositiveabnormalsharereturnconstitutesevidencethatinvestorsonaveragearereactingfavorablytotheexpectedgoodnewsinearnings.Toincreasethepoweroftheinvestigation,theresearchermaywishtosimilarlycompareafewdaysoneithersideofday0.2.4ComparingReturnsandIncomeIfpositiveandnegativeabnormalreturnssurroundinggoodorbadnewsarefoundtoholdacrossasampleoffirms,theresearchermayconcludethatpredictionsbasedonthedecisiontheoryandefficientsecuritiesmarkettheoryaresupported(thatis,accountinginformationisuseful).一種復雜情情況是在公公司公告盈盈余時，公司其他特特定信息也隨之而至至……簡單地把這這類公司從從樣本中剔剔除出去；；另一種復雜雜情況為了了區(qū)分市場場回報和公公司特定回回報，對公公司β的估計……用盈余公告告后一段期期間的數(shù)據(jù)據(jù)估計β，用其他方方法估計β和公司特定定回報，不不區(qū)分市場場回報和公公司特定回回報（Easton&Harris,1991）。并不能保證證市場模型型充分描述述產(chǎn)生股票票收益的實實際過程……Brown&Warner(1980)得出結(jié)論：：對于月回回報窗口，，市場模型型比其他可可選方法表表現(xiàn)得更合合理。01-Jan-2323CurrentFinancialStatementEvidenceGNBNStateHigh0.800.20Low0.100.90（副對角線線概率）削削弱當期財財務報表信信息和未來來公司業(yè)績績之間的關(guān)關(guān)系，稱為為財務報表表中的噪音音（noise）或低盈余余質(zhì)量（lowearningsquality）。主對角角線概率越越高，系統(tǒng)統(tǒng)越有信息息含量（informative），稱為透透明（transparent）或高質(zhì)量量（highquality）信息系統(tǒng)的的信息含量量能夠被實實證檢驗3.FinancialinformationandMarketResponsePermanent:expectedtolastindefinitelyTransitory:affectingearringsinthecurrentyearonlyPriceIrrelevant:zeropersistencyTypesofEarningEvents01-Jan-232401-Jan-2325EventStudyItstudiesthesecuritiesmarketreactiontoaspecificevent.事件研研究是是目前前檢驗驗半強強式有有效市市場假假說的的主要要方法法，用用來了了解資資本市市場證證券價價格與與特定定事件件之間間相關(guān)關(guān)性的的實證證研究究若此事事件有有影響響，證證券價價格波波動狀狀況異異于無無此事事件時時的表表現(xiàn)，，產(chǎn)生生異常常回報報應用統(tǒng)統(tǒng)計方方法檢檢驗異異?；鼗貓鬆顮顩r，，以說說明此此事件件是否否對證證券價價格有有影響響常用事件:公司盈余公公告、新股股發(fā)行、增增發(fā)和配股股、股票回回購或分割割、股利分分配、兼并并收購、盈盈利預測，，以及宏觀觀經(jīng)濟政策策變化公告告等Ball&Brown(BB,1968)study課后自學，，下次課提提問自行設計一一個與BB研究類似研研究構(gòu)想想01-Jan-2326事件及窗口估計期窗口事件期窗口樣本（分組并歸納樣本特征）估計正常和異常收益統(tǒng)計檢驗窗口長短選選擇沒有固固定標準，，但數(shù)據(jù)的的可得性會會制約窗口口長短選擇擇。短窗口口從幾分鐘鐘到幾天，，長窗口可可能涉及幾幾個月到幾幾年短窗口容易易避免事件件窗內(nèi)其他他事件對證證券價格的的影響，但但短窗口可可能錯誤估估計事件窗窗內(nèi)預期收收益率，而而且有些事事件的滯后后影響可能能是短窗口口所不能捕捕獲的。因因此，近年年來長窗口口比較流行行，但長窗窗口也存在在著諸如遺遺漏風險因因素并錯誤誤計量風險險、幸存者者偏差和數(shù)數(shù)據(jù)挖掘偏偏差等數(shù)據(jù)據(jù)問題及統(tǒng)統(tǒng)計推斷問問題正常收益，，假設不發(fā)發(fā)生此事件件的預期收收益，常用用計算模型型：市場模模型、均值值調(diào)整模型型、市場調(diào)調(diào)整模型異常收益，事事件期間內(nèi)證證券實際收益益與同期正常常收益的差01-Jan-2327事件件研研究究法法是是指指運運用用股股票票收收益益率率數(shù)數(shù)據(jù)據(jù)來來測測定定某某一一特特定定經(jīng)經(jīng)濟濟事事件件對對公公司司價價值值的的影影響響。。事事件件研研究究法法先先利利用用估估計計期期,估計計出出事事件件日日的的期期望望收收益益,由事事件件期期的的實實際際收收益益扣扣除除期期望望收收益益得得到到非非正正常常收收益益,再檢檢驗驗樣樣本本平平均均非非正正常常收收益益是是否否顯顯著著區(qū)區(qū)別別于于原原假假設設。。事事件件日日的的期期望望收收益益可可以以由由均均值值調(diào)調(diào)整整模模型型、、市市場場調(diào)調(diào)整整模模型型和和市市場場模模型型來來估估計計。。4.TheBallandBrownStudyBall&Brown(BB,1968)beganatraditionofempiricalmarketsresearchinaccountingthatcontinuestothisday.Theywerethefirsttoprovideconvincingscientificevidencethatfirms’’sharereturnsrespondtotheinformationcontentoffinancialstatements.4.1MethodologyandFindings4.2CausationVersusAssociation4.3OutcomesoftheBBStudy01-Jan-23284.1MethodologyandFindingsBBexaminedasampleof261NYSEfirmsovernineyearsfrom1957to1965.BBconcentratedontheinformationcontentofearnings.BB’sfirsttaskwastomeasuretheinformationcontentofearnings,thatis,goodnews(GN)andbadnews(BN)……….Thus,firmswithearningshigherthanlastyear’’swereclassifiedasGN,andviceversa.Thenexttaskwastoevaluatethemarketreturnonthesharesofthesamplefirmsnearthetimeofeachearningsannouncement.ThiswasdownaccordingtotheabnormalreturnsprocedureillustratedinFigure5.2.TheonlydifferencewasBBusedmonthlyreturns(dailyreturnswerenotavailableondatabasesin1968)BBrepeatedtheirabnormalsecuritymarketreturnscalculationforawidewindowconsistingofeachofthe11monthspriortoand6mothsfollowingthemonthofearningsrelease(month0).01-Jan-23294.1MethodologyandFindingsAveragecumulativeones01-Jan-23304.2CausationVersusAssociationIfasecuritymarketreactiontoaccountinginformationisobservedduringanarrowwindowofafewdayssurroundinganearningsannouncement,itcanbearguedthattheaccountinginformationisthecauseofthemarketreaction.Itcannotbeclaimedthatreportednetincomecausedtheabnormalreturnsduringthe11monthsleadinguptomonth0.Themostthatcanbearguedisthatnetincomeandreturnsareassociated.Wewillfindthattheassociationbetweensharereturnsandearningsincreasedasthewindowwidens(Easton,Harris&Ohlson,1992;Warfield&Wild,1992)01-Jan-23314.3OutcomesoftheBBStudyItopenedupalargenumberofadditionalusefulnessissues:Whetherthemagnitudeofunexpectedearningsisrelatedtothemagnitudeofthesecuritymarketresponse(Beaver,Clarke&Wright,1979).Since1968,accountingresearchershavestudiedsecuritiesmarketresponsetonetincomeonotherstockexchanges,inothercountries,andforquarterlyearningsreports,withsimilarresults.Theapproachhasbeenappliedtostudymarketresponsetotheinformationcontainedinnewaccountingstandards,auditorchanges,etc.Earningsresponsecoefficients(ESC)asksadifferentquestion,namely,foragivenamountofunexpectedearnings,isthesecuritymarketresponsegreaterforsomefirmsthanothers?01-Jan-23325.EarningsResponseCoefficients(ERC)5.1ReasonsforDifferentialMarketResponse5.2ImplicationsofERCResearch5.3MeasuringInvestors’EarningsExpectations5.4SummaryAnearningsresponsecoefficients(ERC)measurestheextentofasecurity’sabnormalmarketreturninresponsetotheunexpectedcomponentofreportedearningsofthefirmissuingthatsecurity.01-Jan-23335.1ReasonsforDifferentialMarketResponseBetaEmpiricalevidenceofalowerERCforhigher-betasecuritieswasfoundbyCollins&Kothari(1989),andbyEaston&Zmijewski(1989).CapitalStructureEmpiricalevidenceofalowerERCformorehighlyleveredfirmswasreportedbyDhaliwal,Lee&Fargher(1991).EarningsQualityThehigherearningsquality,thehigherwewouldexpectedtheERCtobe.Measurementofearningsquality:EarningspersistenceAccrualsqualityOtherreasons01-Jan-2334EarningsQualityDescriptionofeventActualevent01-Jan-2335What’’smeaningofwindow-dressing?EarningspersistenceWewouldexpectthattheERCwillbehigherthemorethegoodorbadnewsincurrentearningsisexpectedtopersistintothefuturefirmperformance.Kormedi&Lipe(1987)Themeasureofpersistencewastheextenttowhichearningschangesofthelasttwoyearscontinuedintothecurrentyear.Ramakrishnan&Thomas(RT,1991)Differentcomponentsofnetincomemayhavedifferentpersistence.Thisimpliesthataccountantsshouldprovidelotsofclassificationanddetailontheincomestatement.Permanent,expectedtopersistindefinitely(ERC>1)Transitory,affectingearningsinthecurrentyearbutnotfutureyears(ERC=1)Price––irrelevant,persistenceofzero(ERC=0)成功功引引進進新新產(chǎn)產(chǎn)品品，，處處置置產(chǎn)產(chǎn)房房和和設設備備，，資資本本化化開開辦辦費費，，注注銷銷研研究究費費………01-Jan-2336HigherearningsqualityHighpersistenceofearningsandcashflowsHighpredictiveabilityofearningsandcashflowsHighearningsresponsecoefficientLowlevelofearningsmanagementMorevoluntarilydisclosureStrongcorporategovernance01-Jan-2337Whatshouldtheusersbeawareof?Statementusersmust:Understandcurrentfinancialreportingsettingsandstandards.Differencesinaccountingmethods.Differencesinaccountingestimates.Differencesinstandardsimplementation.Recognizethatmanagementmaymanipulatethefinancialinformation.Distinguishbetweenreliablefinancialstatementinformationandpoorqualityinformation.3801-Jan-2338AccrualsqualityWewouldexpectthatahigherERCforhigheraccrualsquality.DeChow&Dichev(2002)Earningsqualitydependsprimarilyonthequalityofworkingcapitalaccruals.Totheextentcurrentperiodworkingcapitalaccrualsshowupascashflowsnetperiod,thoseaccrualsareofhighquality.Asimilarargumentappliestolastperiod’’saccruals.Evidencethatfirm’’sERCsandsharepricesrespondpositivelytoaccrualqualityasmeasuredbythisprocedureisreportedbyFrancisetal(2004,2005)andEckeretal(2006).01-Jan-23395.2ImplicationsofERCResearchImprovedunderstandingofmarketresponsesuggestswaysthataccountantscanfurtherimprovethedecisionusefulnessoffinancialstatements:Lowerinformativenessofpriceforsmallerfirmsimpliesthatexpandeddisclosureforthesesfirmswouldbeusefulforinvestors,contrarytoacommonargumentthat…………toexpanddisclosureofthenatureandmagnitudeoffinancialinstruments,includingthosethatare““off-balance-sheet””.……thedesirabilityofdisclosureofsegmentinformation,since………Also,MD&Aenablesthefirmtocommunicateitsgrowthprospect.Disclosureofthecomponentsofnetincomeisusefulforinvestors.01-Jan-23405.3MeasuringInvestors’EarningsExpectationsUndertheidealconditions,expectedearningsissimplyaccretionofdiscountonopeningfirmvalue.Whenconditionsarenotideal:Oneapproachistoprojectthetimeseriesformedbythefirm’spastreportednetincomes,thatis,tobasefutureexpectationsonpastperformance.However,dependsonearningspersistence.Ifearningsare100%persistent,expectedearningsarejustlastyear’sactualearnings,thenunexpectedearningsarethechange(Ball&Brown,1968);Ifearningsare0persistent,unexpectedearningsareequaltothelevelofcurrentyear’’searnings(BillCautions,Example3.1);Easton&Harris(1991)foundbothchangesinandlevelsofnetincomearecomponentsofthemarket’searningsexpectation.(Tobecontinued……)01-Jan-23415.3MeasuringInvestors’EarningsExpectationsUndertheidealconditions,expectedearningsissimplyaccretionofdiscountonopeningfirmvalue.Whenconditionsarenotideal:Anothersourceofearningsexpectationsisanalysts’forecasts.Sincerationalinvestorswillpresumablyusethemostaccurateforecast.Analysts’forecastsaremoreaccuratethantimeseriesforecasts(Brownetal,1987;O’Brien,1988);Thesinglemostrecentearningsforecastprovidedamoreaccurateearningspredictionthantheaverageforecastofallanalystsfollowingthefirm(O’’Brien,1988);Analysts’forecastsareoptimistically,althoughthebiasmayhencedecreasedinrecentyears(Kothari,2001).01-Jan-23425.4SummaryTheinformationcontentofnetincomecanbemeasuredbytheextentofsecuritypricechangeor,morespecifically,bythesizeofthesecurity’’sabnormalmarketreturn,aroundthetimethemarketlearnsthecurrentnetincome.Foragivenamountofunexpectednetincome,theextentofsecuritypricechangeorabnormalreturnsdependsonfactorssuchas…………Theempiricalresultsarereallyquiteremarkable.First,theyhaveovercomesubstantialstatisticalandexperimentaldesignpreambles;Second,theysupportsthetheoryofsecuritiesmarketefficiencyandthedecisiontheoriesthatunderlieit.Finally,theysupportthedecisionusefulnessapproachtofinancialreporting.01-Jan-23436.ACaveatAboutthe““Best””AccountingPolicyHowever,whataccountantscannotdoisclaimthatthebestaccountingpolicyistheonethatproducesthegreatestmarketresponse.Whynot?Thereasonisthatinformationhascharacteristicsofapublicgood.Asaresult,investorsmayperceiveaccountinginformationasusefuleventhoughfromsociety’’sstandpointthecostsofthisinformationoutweighthebenefitstoinvestors.Itisstilltruethataccountantscanbeguidedbymarketresponsetomaintainandimprovetheircompetitivepositionassupplierstothemarketplaceforinformation.Itisalsotruethatsecuritiesmarketswillworkbettertotheextentsecuritypricesprovidegoodindicationsofunderlyingrealinvestmentopportunities.However,thesesocialconsiderationsdosuggestthat,asageneralrule,accountingstandardsettingbodiesshouldbewaryofusingsecuritiesmarketresponsetoguidetheirdecisions.01-Jan-23447.TheInformationContentofOtherFinancialStatementInformationEvidenceofusefulnessismixed:Magliolo(1