12Does Earnings Management Affect Firms’Investment

DoesEarningsManagementAffectFirms’InvestmentDecisions?MaureenF.McNicholsStanfordUniversityStephenR.StubbenTheUniversityofNorthCarolinaatChapelHillABSTRACT:Thispaperexamineswhetherfirmsmanipulatingtheirreportedfinancialresultsmakesuboptimalinvestmentdecisions.Weexaminefixedassetinvestmentsforalargesampleofpubliccompaniesduringthe1978–2002periodanddocumentthatfirmsthatmanipulatetheirearnings—firmsinvestigatedbytheSECforaccountingirregularities,firmssuedbytheirshareholdersforimproperaccounting,andfirmsthatrestatedfinancialstatements—over-investsubstantiallyduringthemisreportingperiod.Furthermore,followingthemisreportingperiod,thesefirmsnolongerover-invest,con-sistentwithcorrectedinformationleadingtomoreefficientinvestmentlevels.Wefindsimilarpatternsforfirmswithhighdiscretionaryrevenuesoraccruals.Ourfindingssuggestthatearningsmanagement,whichislargelyviewedastargetingpartiesexter-naltothefirm,canalsoinfluenceinternaldecisions.Keywords:earningsmanagement;investment.I.INTRODUCTIONThispaperexamineswhetherearningsmanagementaffectsresourceallocationbystudyingwhetherfirmsmanipulatingearningsmakesuboptimalinvestmentdecisions.Weaimtoprovideevidenceonwhetheraccountingmisstatements,potentiallymotivatedbycompen-sationtargetsorcapitalmarketexpectations,causedistortionsintheinvestmentdecisionsmadewithinfirmsengaginginthemisstatement—adirectcosttoinvestors.Theconsequencesofearn-ingsmanagementarerelevanttothedecisionsmadebyinvestors,managers,directors,andregu-lators,andtodatethereislittleresearchthataddressestherelationbetweenearningsmanagementinexternalreportinganditseffectoninternaldecisions.Investmentdecisionsdependonexpectationsofthebenefitsoftheinvestment,whichinturndependonexpectationsoffuturegrowthandproductdemand.ExpectationsoffuturegrowthareTheauthorsthankDanDhaliwal,twoanonymousreviewers,andworkshopparticipantsatthe2005StanfordSummerCamp,the2005UNC/DukeFallCamp,LancasterUniversity,andthe2007AccountingResearchConferenceattheUniversityofOklahomaforinsightfulcommentsandsuggestions.WearegratefultoWoodruff-Sawyer&Co.fortheuseofdataonshareholdersuitsforaccountingimproprieties.WealsogratefullyacknowledgethefinancialsupportoftheStanfordGraduateSchoolofBusinessandtheKenan-FlaglerBusinessSchool.Editor’snote:ThispaperwasacceptedbyDanDhaliwal.THEACCOUNTINGREVIEWAmericanAccountingAssociationVol.83,No.6DOI:10.2308/accr.2008.83.6.15712008pp.1571–1603Submitted:August2007Accepted:May2008PublishedOnline:December20081571basedoninformationthatincludesrevenuesandearnings.Inadditiontomerelyconcealingtheactualperformanceduringtheperiod,misstatedfinancialresultscanmaskunderlyingtrendsinrevenueandearningsgrowth.Thus,overstatementsofrevenuesandearningsarelikelytodistortexpectationsofgrowthbythoseunawareofthemisstatement.Onemightconjecturethatifmanagementchoosestopaintarosierviewforinvestorsinthenumberstheyreportexternally,thentheywouldnotallowthistoinfluenceinternalinvestmentdecisions.However,itispossiblethatinvestmentdecisionmakerswithinthefirmbelievethemisreportedgrowthtrend—becausetheyareeitherover-optimisticorunawareofthemisstatement—andinvestaccordingly.Alternatively,investmentdecisionmakersmightunder-standthetruestateofthefirmbutchoosetoover-investinahigh-riskapproachtoturnaroundperformance.Regardlessofthereasonfortheover-investment,truthfulreportingmighthavepreventedit.Severalpartiesaretypicallyinvolvedininvestmentdecisions,includingmanagerswhomakethedecisiontoinvest,boardswhoreviewthecapitalbudget,andexternalsuppliersofcapital.Iffinancialresultsarereportedtruthfully,thenotherpartiescouldstepintocurtailtheinvestment.Asaresult,firmsinvestmorethantheyotherwisewouldhave,andattemptstomeetcapitalmarketexpectationsormeetbonustargets,forexample,couldaffectinvestors,employees,customers,andabroadsetofrelatedparties.Ourstudyprovidesevidenceonwhetherearningsmanagementaffectsresourceallocationbyexaminingthecapitalexpendituredecisionsofthreegroupsoffirmsallegedtohavemanipulatedearnings—firmsinvestigatedbytheSECforaccountingirregularities,firmssuedbytheirshare-holdersforimproperaccounting,andfirmsthatrestatedfinancialstatements.Thesesamplescom-prisefirmsthathaveoverstatedearningssubstantially,whichgivesusgreaterconfidencethatwehaveidentifiedearningsmanipulatorsandthatthemagnitudeoftheirearningsmanipulationcouldsignificantlyaffectinvestmentdecisions.Bothofthesefactorsarecrucialtothepowerofourtests.However,becauseofconcernsaboutpotentialselectionbiasandtotestwhethertheeffectwefindgeneralizesbeyondthemostseverecasesofearningsmanipulation,wealsousediscretionaryrevenuestoidentifyfirmsthatmanipulatedearnings.Ourdataidentifiestheperiodforwhichmanipulationisalleged,andpermitsustoexaminetheinvestmentdecisionsintheyearsbefore,during,andsubsequenttothemanipulation.Underthenullhypothesisthatearningsmanagementdoesnotresultinresourcemisallocation,firmsshouldexhibitinvestmentlevelsconsistentwiththeirfundamentals.Ourstudyfollowsanextensiveliteraturethatmodelsinvestmentdecisionsusingalinearmodelrelatingcapitalexpen-diturestoinvestmentopportunities.Specifically,thesemodelscontrolforinvestmentopportunitie