会计选择的实证研究外文翻译.docx

会计选择的实证研究外文翻译.docx

《会计选择的实证研究外文翻译.docx》由会员分享，可在线阅读，更多相关《会计选择的实证研究外文翻译.docx（11页珍藏版）》请在冰点文库上搜索。

会计选择的实证研究外文翻译

外文文献翻译译文

原文：

EmpiricalResearchonAccountingChoice

1.Earningsmanagement

BurgstahlerandDichev（1997）reportthatmanagersapparentlymanageearningstoavoidearningsdecreasesandlosses.Theyrelyonatransactionscosttheoryratherthanefficientcontractingormanagerialopportunismtoexplaintheirresults.Thatis,theysuggesttermsoftransactionswithstakeholdersaremorefavorableforfirmswithhigherratherthanlowerearnings（seeBowen,etal.,1995forfurtherdiscussionofthispoint）andalsothatinvestorsarenotfullyrationalinassessingtheinformationcontentofreportedearnings,consistentwithprospecttheory.

Theabovestudiesallreportevidenceofearningsmanagementviachoicesofaccountingmethodsbutnonedocumentanyassociatedpricereactionstothesechoices.Inotherwords,thesestudiesdonotexplorewhethertheseaccountingchoiceshaveeconomicimplications.Barth,etal.（1999）,ontheotherhand,findthatfirmswithatimeseriesofincreasingearningshavehigherpriceearningsmultiplesaftercontrollingforriskandgrowth,thanfirmswithoutanincreasingearningspattern.Thisevidenceisconsistentwiththesuccessofearningsmanagement.however,Barthetal.donotexplicitlytestforearningsmanagementanddonotattributetheearningspatternasnecessarilyduetoearningsmanagement.Davis（1990）,inapartialreplicationandextensionoftheHong,etal.（1978）studyofthepurchaseandpoolingchoice,findsthatacquiringfirmsthatusethepurchasemethodenjoypositiveabnormalreturnsovertheperiodextendingfrombeforetheannouncementofthebusinesscombinationtoafteritsconsummation.Acquiringfirmsusingthepoolingmethodenjoyonlynormalmarketreturns.HisresultsareconsistentwiththoseofHong,etal.

2.Multiplemotivations

Inadditiontotheproblemofaddressingmultipleaccountingchoices,generallyasreflectedinaccruals,thereisalsotheissueofmultiple,andpotentiallyconflicting,motivationsfortheaccountingchoices.Mostoftheworkdiscussedinsection4focusesonasinglemotiveforaccountingchoicedecisions.Forexample,thecompensationliteraturefocusesonthequestionofwhethermanagersuseaccountingdiscretiontomaximizetheircompensation.Implicitly,theresultssuggestthatmanagers’actionscomeattheexpenseofshareholders.Butifthisisso,whydocompensationcontractsallowdiscretion?

Oneplausibleansweristhatmanagers’actionsarenotonlyanticipated,butalsodesirablefromshareholders’perspective.Forexample,thesameaccountingchoicesthatmaximizemanagers’compensationmayalsodecreasebondcovenantviolationsorincreaseassetvaluations.However,suchmotivesaretypicallynotincludedintheanalysis.Byfocusingononegoalatatime,muchoftheliteraturemissesthemoreinterestingquestionoftheinteractionsbetweenandtrade-offsamonggoals.Moreover,itisnotclearwhethertheconclusionsareattributabletothespecificmotivationbeinganalyzed;generallyresultsconsistentwithonehypothesisareconsistentwithmany.Forexample,whatmayappeartobeanopportunisticchoiceofanearningsincreasingaccountingmethodchoice（tobenefitmanagersattheexpenseofotherstakeholdersinthefirm）,maybeinfactaresponsetoavoidabondcovenantviolation（andthusbenefitsallotherstakeholdersattheexpenseofthecreditors）.Finally,withonlyfewexceptions,researchinthe1990sgenerallyfocusesonmotivesidentifiedinthe1970sand1980s–typicallytheusualsuspectsareroundedup.However,wesuspectthatnewinsightsmaybegainedbyinvestigatingadditionalmotives.

Theproblemofmultipleconflictscanbeviewed,inturn,asaspecialcaseofthefamiliar“correlatedomittedvariable”problemineconometrics.Forexample,iffirmswithcompensationcontractsthatdependonearningsarealsomorelikelytohavehighpoliticalcosts,thenacompensationstudycouldfindarelationshipbetweencompensationcontractsandaccountingchoice,eveniftheaccountingchoicesareinfactdrivenbypoliticalconsiderations.Theusualsolutiontothisproblemistoaddcontrolvariables.However,thissolutionasitisappliedtoaccountingchoiceresearchsuffersfromatleastthreedrawbacks.First,researchersoftenrelyoncoarseorinappropriateproxiestomeasuretheroleoftheomitteddeterminantsofaccountingchoice.Forexample,incompensationstudies,leverageandsizevariableshavebeenusedtoproxyforbondcovenanteffectsandpoliticalcosts.

Second,inferenceproblemsarelikelytoarisewhenanalyzingmultiplemotivationsusingproxieswithdifferingamountsofmeasurementerror,particularlywhentheunderlyingeffectsarecorrelated.Insuchinstances,theproxywiththeleastamountofmeasurementerrorislikelytodominate,evenwhenitstrueeffectmaynotbethemostimportant.Forexample,assumethatpoliticalcostsandcompensationmotivationsarecorrelated,butthataccountingchoicesaredrivenbypoliticalcosts.Nevertheless,ifsizeisusedasa（noisy）proxyforpoliticalcostswhiletheresearcherisabletomeasurethedesignofthecompensationcontractwithgreatprecision,thenaregressionofaccountingchoiceonsizeandcompensationcontractdesignwillloadonthemoreprecise（butlesseconomicallyimportant）compensationvariable,ratherthantheeconomicallysignificant（butnoisy）proxyforpoliticalcosts.Little,ifany,progresshasbeenmadeinaddressingtheseconcerns.

Finally,theproblemofmultiplemotivationsisfurthercompoundedbytheabsenceoflinearity.Specifically,discussionofthecorrelatedomittedvariableissueoften（atleastimplicitly）assumesthatthevariables（boththoseincludedandthoseomitted）arelinearandthattheomittedvariableisadditivelyseparablefromthevariableofinterest.Theseassumptionsaremadetofacilitateresearchdesign.However,thereisnoevidencethatsuchassumptionsarejustified.

Eventhosepapersthatconsidermultiplemotivationsgenerallytreatthemindependently.Inpractice,ofcourse,managersfaceavarietyofconflicts,whichwillnot,ingeneral,suggestconsistentcoursesofactions.Underthesecircumstances,firmsmustmaketrade-offsamongthevariousgoals.

3.Conclusionandsuggestionsforfuture

Wedonotwanttoleavetheimpressionthatresearchershavegainednoknowledgeoftheroleandtheimportanceofaccountingchoice.Rather,ourconcernisthatprogresshasslowed.Inpart,thisisduetounambitiousattemptstoexpandthefield.Forexample,testingtheimplicationsofonemoreaccountingstandardaddsverylittletothecumulativeknowledge.Amoreintransigentproblemisthedifficultyinspecifyingresearchdesignsthataccommodatethecomplexityofthetaskathand:

thatis,thesimultaneousimpactofmultiplechoices,multiplegoalsandeconometriccomplications.Ratherthancontinuetoreplicatewell-knownresultsinslightlydifferentsettings,wefeelthatitisimportantforresearcherstograpplewiththesemoredifficult,andatheart,morefundamentalissues.Wehavethreespecificrecommendationsforfutureresearch.

First,researchresultsfailtoprovidecompellingevidenceoftheimplicationsofalternativeaccountingmethodsandwerecommendmoreeffortstodeterminethenatureofsuchimplications.Theliteratureprovidesamplecircumstantialevidencethataccountingchoicemattersbutlittledirectevidence.Forexample,theliteraturedocumentsthatmanagersmakeaccountingchoicesconsistentwithbonusmaximization,butdoesnotdeterminewhetherthisbehaviorresultsinincreasedcashpayouts.Evenifsuchevidencewereproduced,thenextstepwouldbetoascertainwhetherexpectedtotalcompensationincreasesandwhetherthisresultwasintendedand/oranticipatedbythecontractingparties.Thatis,ifaccountingchoicehassuchpotentialwealthimplications,thencontractingpartiesshouldpriceaccountingchoice.One（modest）attemptatthisissuewasmadebyHealy,Kang,andPalepu（1987）forthecaseofmanagerialcompensation,butmoreisneeded.Anotheravenueistoinvestigatethecostscompaniesarewillingtoincurtomaintainaccountingmethodchoicediscretion.AfirstattemptatthisapproachcanbefoundinBeatty,etal.（2000）,andweencouragemoreexplorationofthisissue.

Similarly,thereisnoconsistentevidencesupportingclaimedvaluationdifferencesduetoaccountingmethods.Wedonotknowwhetherthisisbecauseaccountingdifferencesdonotaffectfirmvaluationorbecauseempiricalmethodsareinadequatetodetectanysucheffect.Althoughmuchexistingresearchistiedtostockreturns,accountingnumbersgenerallyexplainonlyasmallportionofthevariabilityinstockreturns,raisingthequestionofwhetherthisistheappropriateplacetolookforaneffect.Despiteextensiveacademicresearch,theprocessbywhichsecuritypricesareset,includingtheinfluenceofaccountingdata,remainsunknown.

Second,becauseaccountingisusedformanypurposes,wehavearguedthatitisinappropriatetoanalyzeoneaccountingissueorevenonegoalinisolation.Ideally,onewouldhaveacomprehensivetheoryofaccountingchoice,butsuchatheoryispresentlynotavailable,anditsdevelopmentdoesnotappearimminentbecauseofthecomplexitiesinherentinsuchamodel.Analyticalmodelscanhelpprovideguidancetoresearchersinstructuringtheempiricalexperiments,inidentifyingappropriatevariables,andinformulatingalternativehypotheses.Atpresent,mostanalyticalmodelsaresoabstractastoofferonlylimitedguidancetoempiricists.Wedonotwanttogivetheimpressionthatweunderestimatethecomplexityofthistask.Furthermore,wearecautiouslyoptimisticthatprogresstowardthenecessarytheoreticali