The-Conservatism-Principle-and-the-Asymmetric-Time

JOURNALOFAccounting&EconomidsELSEVIERJournalofAccountingandEconomics24(1997)337Theconservatismprincipleandtheasymmetrictimelinessofearnings1SudiptaBasu*DepartmentofAccountancy,BaruchCollege,CUNY,17LexingtonAvenue,NewYork,NY1001~Received1May1995;receivedinrevisedform1June1997USAAbstractIinterpretconservatismasresultinginearningsreflecting'badnews'morequicklythan'goodnews'.Thisinterpretationimpliessystematicdifferencesbetweenbadnewsandgoodnewsperiodsinthetimelinessandpersistenceofearnings.Usingfirms'stockreturnstomeasurenews,thecontemporaneoussensitivityofearningstonegativereturnsistwotosixtimesthatofearningstopositivereturns.Ialsopredictandfindthatnegativeearningschangesarelesspersistentthanpositiveearningschanges.Earningsresponsecoefficients(ERCs)arehigherforpositiveearningschangesthanfornegativeearningschanges,consistentwiththisasymmetricpersistence.©1997ElsevierScienceB.V.Allrightsreserved.JELclassification:M41;D21;D82;G14;G38;K41;N20Kevwords:Capitalmarkets;Earningsreturnrelation;Persistence;Auditorlegalliability;Negativereturns*Correspondingauthor.Tel.:[phone];fax:[phone];e-mail:[e-mail]issertationcompletedattheUniversityofRochester.1amgratefulfortheadviceandencouragementofmycommittee,RayBall(Chairman),RossWattsandS.P.Kothari.MikeBarclay,AndrewChristie,MicheleDaley,SteveLilien,NellPearson,TerryShevlin(thediscussant),JoeWeintrop,JerryZimmerman(theeditor),andtwoanonymousrefereeshavealsocontributedsignificantlytothispaper,lalsothankworkshopparticipantsattheUni-versityofArizona,UniversityofBuffalo,BaruchCollege,McGillUniversity,NewYorkUniversity,the1995JAEConference,the1996ConferenceonFinancialEconomicsandAccounting,andespecially,Ph.D.workshopparticipantsattheSimonSchoolforhelpfulcomments.Allerrorsaretheauthor's.0165-4101/97/$17.00~.~)1997ElsevierScienceB.V.Allrightsreserved.PHS0165-4101(97)00014-14S.Basu/JournalofAccountingandEconomics24(1997)3371.IntroductionThispaperre-examinestheconservatismprinciple.Iinterpretconservatismascapturingaccountants'tendencytorequireahigherdegreeofverificationforrecognizinggoodnewsthanbadnewsinfinancialstatements.Undermyinterpretationofconservatism,earningsreflectsbadnewsmorequicklythangoodnews.Forinstance,unrealizedlossesaretypicallyrecognizedearlierthanunrealizedgains.Thisasymmetryinrecognitionleadstosystematicdifferencesbetweenbadnewsandgoodnewsperiodsinthetimelinessandpersistenceofearnings.Fig.1illustratesthepredictionstestedinthispaper.Considerafirmreceivingnewsthatchangesitsestimateoftheproductivelifeofafixedasset.PanelAdepictstheeffectsofthenewsonthereportedbookvalueoftheasset,andPanelBdepictstheassociatedeffectonthefirm'searnings.Ifthenewestimatedlifeislonger,thefirmiseconomicallybetteroff,butunderhistoricalcostaccountingnogainisrecordedcurrently.Instead,thedepreciationchargesthatPanelA70,000[.Originalestimate60,000-~................/-l-Longerestimate___N~I+~Sho_~erestimate50000..........,o.000..............,¢,................-~~x,tt,,,=o,ooo--\...................I0,000.......0IIIII~-[I•'r=012345678910111213Year--endFig.1.PanelA:Exampleofbookvaluesofafixedassetunderconservativeaccountingwhenestimatesofremainingusefullifechange.Straightlinedepreciationwithzerosalvageovertenyearsoriginally.Attheendofthefourthyear,theexpectedlifeisincreasedordecreasedbythreeyears.Write-offassumedtobethreeyears'depreciation.S.Basu/JournalofAccountingattdEconomics24(1997)3-375PanelB40,00035,00030,00025,0008z20,00015,ooo.10,0005,000[IIrIq[IIIII12345678910111213Year-endFig.t.PanelB:Exampleofreportednetincomeunderconservativeaccountingwhenestimatesofremainingusefullifeofafixedassetchange.Straightlinedepreciationwithzerosalvageovertenyearsoriginally.Attheendofthefourthyear,theexpectedlifeisincreasedordecreasedbythreeyears.Write-offassumedtobethreeyears'depreciation.Revenuesandotherexpensesassumedtobe$41,000ineachyearofthefixedasset'slife.wouldhavebeentakeninthecurrentandfutureperiodsarespreadoutoverthenewremaininglife,resultinginlowerdepreciationchargesinPanelA,andhigherincomeinPanelB,eachyear.2Iftheexpectedlifedecreases,asymmetrictreatmentwouldincreasedepreciationchargesovertheentireshorterremaininglife.Inpractice,however,theaccountantrecordsanassetimpairment(PanelA),whichresultsin(oftensharply)reducedcurrentincome,butnoeffectonfutureincomeinPanelB.Inshort,reportedearningsrespondsmorecompletelyorquicklytobadnewsthangoodnews.Myfirstpredictionisthatearningsismoretimelyorconcur-rentlysensitiveinreflectingpubliclyavailable'badnews'than'goodnews'.Totestthispredictionofasymmetrictimeliness,lusenegativeandpositive2Eventhoughincomeislowerbythenewdepreciationchargeinthelastyearsofthenewestimatedlife,netincomeishigherbecausethefixedassetcontributesrevenuesintheseyears.6S.Basu/JournalofAccountingandEconomics24(1997)3-37unexpectedannualstockreturnstoproxyfor~badnews'and'goodnews',respectively.Mysecondpredictionisthattheconcurrentearnings-returnassociationisrelativelystrongerthantheconcurrentcashflow-returnassociationforpubliclyavailable'badnews'comparedto'goodn