Debt-Covenant-Violation-and-Manipulation-of-Accrua

JournalofAccountingandEconomicsI7(1994)145-176.North-HollandDebtcovenantviolationandmanipulationofaccruals*MarkL.DeFondUniversilyofSouthernCalifornia,LosAngeles,CA90089,USAJamesJiambalvoUniversityofWashington,Seattle.WA98195,USAReceivedJuly1990,finalversionreceivedNovember1992Thispaperexaminestheabnormalaccrualsofasampleof94firmsthatreporteddebtcovenantviolationsinannualreports.Weexpectdebtcovenantrestrictionstoinfluenceaccountingchoicesintheyearprecedingandtheyearofviolation.Time-seriesandcross-sectionalmodelsareusedtoestimate‘normal’accruals.Intheyearpriortoviolation,bothmodelsindicatethat‘abnormal’totalandworkingcapitalaccrualsaresignificantlypositive.Intheyearofviolation,thereisevidenceofpositiveabnormalworkingcapitalaccrualsaftercontrollingformanagementchangesandauditorgoingconcernqualifications.Keywords:Contracting;Accountingchoice;Debtcovenantviolation;Accrualsmanipulation1.IntroductionPositiveaccountingtheorypredictsthatfirmsapproachingcovenantviola-tionwillmakeincome-increasingaccountingchoicestoloosentheirdebtconstraints[WattsandZimmerman(1986)l.lThispredictionistestedbyCorrespondenceto:MarkL.DeFond,SchoolofAccounting,UniversityofSouthernCalifornia,LosAngeles,CA90089,USA.*WeappreciatethecommentsofHelenAdams,PhilBeaulieu,GaryBiddle,WaltBlacconiere,BobBowen,DaveBurgstahler,PaulCaster,LindaDeAngelo,DanDhaliwal,MikeDuffy,MikeEttredge,PatHughes,LaurenKelly,LindaMcDaniel,SusanMoyer,BethRose,ChrisStinson,BobTrezevant,MarkTrombley,BillWaller,andespeciallyJimManegold,AmySweeney(thereferee),andRossWatts(theeditor).ThepaperalsobenefitedfromcommentsbyseminarparticipantsattheUniversityofArizona,USC,andtheUniversityofWashington.ResearchassistancewasprovidedbySteveGlover,SteveMatsunaga,RoseMartin,andMikeStaheli.‘SeeBegley(1990)forareviewofthreerecentstudiesofdebtcovenantsandaccountingchoice.And,seeHolthausen(1990)foradiscussionthatcontrastsopportunisticexplanationsforaccountingchoice(suchasmakingaccountingchoicestoloosendebtconstraints)toefficientcontractingandinformationperspectives.0165-4101/94/[e-mail]abnormalaccrualsexaminingtheabnormaltotalandworkingcapitalaccrualsofasampleof94firmsthatreportedcovenantviolationsinannualreportsduringfiscalyears1985through1988.’Weassumethatifcovenantrestrictionsinfluenceaccount-ingchoice,theinfluencewillbemanifestedintheyearprecedingandtheyearofviolationandouranalysesfocusontheseperiods3‘Normal’accrualsareestimatedusingtime-seriesandcross-sectionalmodels.Results,usingbothmethodsofestimatingnormalaccruals,indicatethatintheyearpriortocovenantviolation,‘abnormal’accrualsaresignificantlypositive.Thisfindingsupportstheconventionalviewthatdebtagreementsmotivatemanagerstomanipulateincome.Resultsintheyearofviolationaremoredifficulttointerpretthanresultsintheyearpriortoviolation.Thisisduetosampleselectionbias-thefirmsinoursampleareknowntohaveviolateddebtcovenants.Firmsthat‘successfully’avoidedviolationbymanipulationareomittedfromthesample.Thus,violationfirmsarelesslikelytoreportpositiveabnormalaccrualsintheyearofviolationbecausethereisno‘reasonable’amountofmanipulationthatwouldallowthemtoavoidviolation.Inspiteoftheselectionbias,thereisstillevidenceofpositivemanipu-lationintheyearofviolationaftercontrollingformanagementchangesandauditorgoingconcernqualifications.Priorresearch[Moore(1973)andDeAngelo(1988)]indicatesthatnewmanagershaveatendencytotakeanearnings‘bath’.Inaddition,Stice(1991)indicatesthatauditorsoffirmsinpoorfinancialconditionfaceincreasedlitigationrisk.Thus,auditorswhoissueagoingconcernqualification(indicatingsubstantialdoubtthatthefirmwillcontinueinexistenceandahigherlikelihoodthattheauditorswillbesued)arelikelytoincreasemonitoringofmanagers’accountingchoicesandinsistonconservativeaccountingpractices(leadingtotheobservationofnegativeabnormalaccruals).Aftereliminatingfirmswithmanagementchangesorauditorgoingconcernqualifications,abnormaltotalaccrualsarepositivebutnotsignificantlydifferentfromzero.Abnormalworkingcapitalaccrualsaresignificantlypositive.Thus,inspiteofasampleselectionbiasagainstdetectionofmanipulation,thereisevidenceofpositivemanipulationintheyearofviolation.Theremainderofourpaperisorganizedasfollows.Inthenextsectionwebrieflyreviewrelatedstudies.Section3presentssampleselectionprocedures‘Throughoutthepaper,werefertothesefirmsasviolationfirms.3Whileweexpectthattheyearofandtheyearprecedingviolationareperiodswhereincentivestomanipulateaccrualsarelikelytobehigh,debtcovenantsmayalsoleadtomanipulationinearlierandlaterperiods.Forexample,butformanipulation,firmsmayhaveviolateddebtcovenantsinseveralperiodspriortotheirinitialviolation.Inaddition,Sweeney(1992)pointsoutthatdebtcovenantviolationsoftencontinueforseveralyearswithincreasingdefaultcosts.Thissuggeststhatmanagersmaycontinuetomanipulateaccrualsinperiodssubsequenttotheinitialviolationinordertoavoidthesecosts.M.L.DeFondandJ.Jiambalvo,Debtcovenantviolationandabnormalaccruals147andaprofileofsamplefirms.Section4presentsti