Evidence-on-the-Trade-Off-between-Real-Activities-

Electroniccopyavailableat:=1891759Electroniccopyavailableat:=18917590 Evidenceonthetrade-offbetweenrealactivitiesmanipulationandaccrual-basedearningsmanagementAmyY.ZangTheHongKongUniversityofScienceandTechnologyAbstract:Istudywhethermanagersuserealactivitiesmanipulationandaccrual-basedearningsmanagementassubstitutesinmanagingearnings.Ifindthatmanagerstradeoffthetwoearningsmanagementmethodsbasedontheirrelativecostsandthatmanagersadjustthelevelofaccrual-basedearningsmanagementaccordingtothelevelofrealactivitiesmanipulationrealized.Usinganempiricalmodelthatincorporatesthecostsassociatedwiththetwoearningsmanagementmethodsandcapturesmanagers’sequentialdecisions,Idocumentlargesampleevidenceconsistentwithmanagersusingrealactivitiesmanipulationandaccrual-basedearningsmanagementassubstitutes.Keywords:realactivitiesmanipulation,accrual-basedearningsmanagement,trade-offDataAvailability:Dataareavailablefrompublicsourcesindicatedinthetext.Iamgratefulfortheguidancefrommydissertationcommitteemembers,JenniferFrancis(chair),QiChen,DhananjayNanda,PerOlssonandHanHong.IamalsogratefulforthesuggestionsandguidancereceivedfromStevenKachelmeier(senioreditor),DanDhaliwalandtwoanonymousreviewers.IthankAllenHuang,MosheBareket,YvonneLu,ShivaRajgopal,MohanVenkatachalamandJerryZimmermanforhelpfulcomments.IappreciatethecommentsfromtheworkshopparticipantsatDukeUniversity,UniversityofNotreDame,UniversityofUtah,UniversityofArizona,UniversityofTexasatDallas,DartmouthCollege,UniversityofOregon,GeorgetownUniversity,UniversityofRochester,WashingtonUniversityinSt.LouisandtheHKUST.IgratefullyacknowledgethefinancialsupportfromtheFuquaSchoolofBusinessatDukeUniversity,theDeloitteFoundation,UniversityofRochesterandtheHKUST.Errorsandomissionsaremyresponsibility.Electroniccopyavailableat:=1891759Electroniccopyavailableat:=18917591 I.INTRODUCTIONIstudyhowfirmstradeofftwoearningsmanagementstrategies,realactivitiesmanipulationandaccrual-basedearningsmanagement,usingalargesampleoffirmsover1987–2008.Priorstudieshaveshownevidenceoffirmsalteringrealactivitiestomanageearnings(e.g.,Roychowdhury2006;Grahametal.2005)andevidencethatfirmsmakechoicesbetweenthetwoearningsmanagementstrategies(Cohenetal.2008;CohenandZarowin2010;Badertscher2011).Mystudyextendsresearchonthetrade-offbetweenrealactivitiesmanipulationandaccrual-basedearningsmanagementbydocumentingasetofvariablesthatexplainthecostsofbothrealandaccrualearningsmanagement.Iprovideevidenceforthetrade-offdecisionasafunctionoftherelativecostsofthetwoactivitiesandshowthatthereisdirectsubstitutionbetweenthemafterthefiscalyearendduetotheirsequentialnature.Realactivitiesmanipulationisapurposefulactiontoalterreportedearningsinaparticulardirection,whichisachievedbychangingthetimingorstructuringofanoperation,investmentorfinancingtransaction,andwhichhassuboptimalbusinessconsequences.TheideathatfirmsengageinrealactivitiesmanipulationissupportedbythesurveyevidenceinGrahametal.(2005).1Theyreportthat80percentofsurveyedCFOsstatedthat,inordertodeliverearnings,theywoulddecreaseresearchanddevelopment(R&D),advertisingandmaintenanceexpenditures,while55percentsaidtheywouldpostponeanewproject,bothofwhicharerealactivitiesmanipulation.                                                            1Inparticular,Grahametal.(2005)notethat:“TheopinionofmanyoftheCFOsisthateverycompanywould/shouldtakeactionssuchofthese[realactivitiesmanipulation]todeliverearnings,aslongastherealsacrificesarenottoolargeandaslongastheactionsarewithinGAAP.”Grahametal.furtherconjecturethatCFOs’greateremphasisonrealactivitiesmanipulationratherthanaccrual-basedearningsmanagementmaybeduetotheirreluctancetoadmittoaccounting-basedearningsmanagementintheaftermathoftheEnronandWorldcomaccountingscandals.2 Unlikerealactivitiesmanipulation,whichalterstheexecutionofarealtransactiontakingplaceduringthefiscalyear,accrual-basedearningsmanagementisachievedbychangingtheaccountingmethodsorestimatesusedwhenpresentingagiventransactioninthefinancialstatements.Forexample,changingthedepreciationmethodforfixedassetsandtheestimateforprovisionfordoubtfulaccountscanbiasreportedearningsinaparticulardirectionwithoutchangingtheunderlyingtransactions.Thefocusofthisstudyisonhowmanagerstradeoffrealactivitiesmanipulationandaccrual-basedearningsmanagement.Thisquestionisimportantfortworeasons.First,asmentionedbyFieldsetal.(2001),examiningonlyoneearningsmanagementtechniqueatatimecannotexplaintheoveralleffectofearningsmanagementactivities.Inparticular,ifmanagersuserealactivitiesmanipulationandaccrual-basedearningsmanagementassubstitutesforeachother,examiningeithertypeofearningsmanagementactivitiesinisolationcannotleadtodefinitiveconclusions.Second,bystudyinghowmanagerstradeoffthesetwostrategies,thisstudyshedslightontheeconomicimplicationsofaccountingchoices;thatis,whetherthecoststhatmanagersbearformanipulatingaccrualsaffecttheirdecisionsaboutrealactivitiesmanipulation.Assuch,thequestionhasimplicationsaboutwh