资讯性激励补偿和缓冲库存的选择外文翻译.docx

资讯性激励补偿和缓冲库存的选择外文翻译.docx

《资讯性激励补偿和缓冲库存的选择外文翻译.docx》由会员分享，可在线阅读，更多相关《资讯性激励补偿和缓冲库存的选择外文翻译.docx（6页珍藏版）》请在冰豆网上搜索。

资讯性激励补偿和缓冲库存的选择外文翻译

外文文献翻译译文

一、外文原文

原文：

Informativeness,IncentiveCompensation,andtheChoiceofInventoryBuffer

Baiman,Stanley;Netessine,Serguei;Saouma,Richard

Mostoftheagencytheoryworkinmanagerialaccountinghasstudiedtherelationbetweenperformancemetricpropertiese.g.,informativeness,precision,congruity_andoptimalincentivecompensation,holdingfixedthefirm’sotherorganizationaldesigndecisions_e.g.,hierarchicalstructure,jobassignment,productiontechnology_.However,byholdingfixedtheseotherorganizationaldesignchoices,onecannotstudywhetherandhowtheyaffecttheperformancemetricproperties.AsHemmer_1998,321–322_notes,“thevalueofaperformancemeasureisdeterminednotsimplybyitscongruityandprecisionbutbyitsinfluenceontheoptimalorganizationaldesignMuchoftherecenttheoreticalaccountingliteraturehaslargelyignoredcomplementaritiesbetweenperformancemeasuresandorganizationaldesign.”

Inthisstudy,weexpandonHemmer’s_1998_observationbystudyinghowtheinformativenessandincentivepropertiesofaperformancemetricareinfluencedbyoneofthefirm’sorganizationaldesignchoices—thesizeofitsinventorybuffers.Theintroductionofjust-in-time_JIT_and,moregenerally,leanmanufacturinghasledtoanincreasedemphasisoncontrollingandreducinginventorylevels.Managerialaccountinghasrespondedtothisreductionininventorywithnewcostingtechniquessuchasbackflushcosting_Horngrenetal.2008_.However,weshowthatthechoiceofinventorybuffershasamoresubtleeffectonthedesignofthemanagerialaccountingsystem,inthatitaffectstheinformativenessofperformancemetricsproducedbythemanagerialaccountingsystem.

Similartoourmodel,Hemmer_1995,1998_andGietzmannandHemmer_2002_examinehowdifferentworkflowarrangementsbetweenagentsaffecttheinformationavailableforcontracting,andtheincentivesfacingagents.Ourworkisdistinctfromthesemodelsthatdonotconsiderbuffersizeasoneoftheprincipal’schoicevariables.Nagaretal._2009_examinetheroleofinventorybuffersinagencyproblems,althoughunlikeourmodel,intheirsettingbuffersarefilledbyagentsinordertosignalprivateinformationandbuffersizeisnotachoicevariable.Allesetal._1995_dofocusontheeffectthatthechoiceofbuffersizecanhaveontheinformativenessofperformancemetrics.Themajordifferencebetweenourworkandtheirsisthattheydonotformallymodelhowbuffersaffectperformancemetrics,butinsteadassumeamonotonicrelation.Incontrast,weformallymodeltheinventoryprocessandderiveanon-monotonicrelationbetweenbuffersizeandtheinformativenessoftheagent’sperformancemetric.

Themodelconsistsofasingle-periodandafirmcomprisedofarisk-neutralprincipalandagent.Contractingtakesplaceatthestartoftheperiodwhentheagentishiredtosetupaworkstationtoprocessincomingintermediateunits_e.g.,testcompletedcomputersorweldautomotivechassis_.Theintermediateunitsarrivestochasticallyintotheworkstation’sincominginventorybufferatacommonlyknown,meanarrivalrateof_.Thebufferhascapacityb_N,allowingamaximumofb_1unitstobeheldinfrontoftheworkstationwhiletheworkstationprocessesoneunit.Ifanintermediateunitarrivesandthebufferisnotfull,thentheunitisaddedtotheinventory.However,ifthebufferisfull,thenblockingoccursandintermediateunitsceasearrivinguntilthereisspaceinthebuffer,whereupontheintermediateunitsbeginarrivingatthesamestochasticrateasbefore.

OurassumptionsregardingarrivalandprocessingratesfollowthestandardM/M/1/bqueuingmodelwithfinitebuffersusedintheoperationsmanagementliterature.7Notethatwhilewearemodelingasingle-periodinthattheagentchooseshissetupeffortonlyonceatthestartoftheperiod,weanalyzethestochasticevolutionoftheworkstation’sthroughputovertheentireperiod.Thetransientbehaviorofthroughputinqueuingsystemswithfinitebufferscannotbedescribedanalytically,althoughthesteady-statebehaviorcanbedescribedinclosedform.8Asaresult,theoperationsmanagementliteraturetypicallyusessimulationstoanalyzethetransientbehaviorofqueuingmodelswithfinitebuffersandclosed-formanalysistoexaminesteady-statebehavior.Incontrast,agencymodelshaveemphasizedrewardingtransientbehavior,butsimplifytheproductionprocess_e.g.,usingtheLENmodel_toachieveclosed-formsolutions.9Inordertoincorporatestochasticintermediateunitarrivalrates,stochasticprocessingrates,andfiniteinventorybuffersinouranalysis,wefocusontheproblem’ssteady-state.Thatis,weassumethattheprincipalisinterestedinmaximizinghersteady-stateexpectedprofitandthattheagentselectshisefforttomaximizehissteady-stateexpectedutility.10Thisisareasonableapproximationif,asweassume,theproductionprocessreachessteady-statefairlyquickly.Ourassumptionthattheagent’seffortisexpendedinsettinguptheworkstationbeforeproductionbegins,ratherthanmanagingtheproductionprocessasitevolves,isalsoconsistentwithouremphasisonthesteady-statebehavioroftheworkstation.Anexampleofourmodeledsettingisaroboticlinethatweldsautomotivechassis.Theagent’smaintaskistoprogramthespot-weldingrobots.Thistaskinvolvesidentifyingtheoptimalpositionofthechassis,identifyingtheoptimalpositionoftheroboticarms,andmakingsurethatrobotpathsdonotcrosseachotherorthechassisitself_sothatnothingisdamagedintheprocess_whileensuringthatweldingisdoneintheleastpossiblenumberofsteps.Allofthisplanningandprogrammingsetupisdoneoncebytheagentatthebeginningoftheshiftandthereafter,aslongastheworkstationisincontrol,theagentdoesnotneedtointervene.Clearly,theactionsoftheagentinprogrammingtherobotsaffectthetimeittakestoweldthechassis_onaverage_,buttheagentdoesnotreprogramtherobotonceproductionhasbegun.

Mostofthetimetheyalsohaveahighdegreeofeducationorexpertise.Theyincludeanywherefromaquartertoathirdoftheworkforce,butnoteveryonewhousesknowledge.Ifyouarediggingditches,youmayhavesomeknowledgeonthejob,butit'snottheprimarypurposeofwhatyoudo.

Arecompaniesdoingagoodjobofmanagingandimprovingtheperformanceofknowledgeworkers?

They'renot.WhatmostorganizationsdoisHSPALTA:

Hiresmartpeopleandleavethemalone.We'vespentalotofeffortrecruitingknowledgeworkersandassessinghowcapabletheymightbebeforewehirethem.Butoncethey'rehiredwedon'tdoalottoimprovetheirperformance.Processimprovementhasmostlybeenforotherworkers:

transactionalworkers,manufacturingworkers,peopleincallcenters.Alltheseriousapproachestoimprovingworkhavelargelyescapedknowledgework.

Weletknowledgeworkersgetawaywithsayingthere'snoprocesstotheirwork,thateverydayisdifferent.Wedon'tmeasuremuchofanythingaboutknowledgework.

Ifwedon'tmeasureknowledgework,whydoyouthinkthere'sroomtoimproveknowledgeworkerproductivityandperformance?

It'saprettywell-informedhunch.Peopleimproveprocessesallthetime;theyjusthaven'tdoneitwithknowledge-workprocessesasmuch.It'sanextrapolationofthesamelogicinotherwork,thatprocessescanbeimproved.

Hereisonenumberthatindicatesperformanceandproductivitycanbeimproved:

IDCfoundthat1,000knowledgeworkerscanloseasmuchas$6millionayearjustsearchingfornonexistentdata,orrepeatingworkthathasalreadybeendone.Isitpossibleeveryknowledgeworkerisworkingtohisorherpotential?

It'spossible,butunlikely.Wecangetalotbetteratimprovingtheirperformance.

Whyhasn'tknowledgemanagementhelpedmoreintheefforttoimproveknowledge-workerperformanceandproductivity?

Knowledgemanagementwasanearlyattempttointerveneinknowledgework.Forthemostpart,itwasn'tparticularlysuccessful,becausewedidn'tlookcloselyathowknowledgeworkersdidtheirwork.Wetriedtobetoobroadinourfocus.

Mostorganizationssimplycreatedonebigrepositoryforallknowledgeandallworkers.

Theonlywaywecangetpeopletouseknowledgeonthejobistounderstandhowtheydotheirjobs,andthenfigureoutsomewaytoinjectknowledgeintothecourseoftheirday-to-daywork,notmakeitaseparatethingyouhavetoconsultwhenyouneedknowledge.

Wehavetobemuchmoretargetedinapproachingknowledgemanagement.Wehavetotargetaspecificjob.Andthebestwayistousetechnologytobaketheknowledgeintothejob.

Notethatthemeanarrivalrateofincomingintermediateunitsandthereforethemaximummeanthroughputis_.However,iftheprincipalinducedtheagenttochooser__/_1_h_sothattheworkstationwassettoprocesstheintermediateunitsatthesameaveragerateatwhichtheyarrived,thentheinstantaneousprobabilityofstarvingtheworkstationwouldbeps=11+b_0,whichinturnwouldlimittheworkstation’smeansteady-statethroughputtoK=_b1+b_.Thatis,inducingtheagenttosettheworkstationtoprocesstheintermediateunitsatthesameaveragerate

atwhichtheyarrivewouldsignificantlylimitsteady-statethroughputandinduceanon-trivialprobabilityofstarving.Thus,theprincipalhasanaturalincentivetoinducetheagenttosettheworkstationtoprocessunitsatanaverageratethatisstrictlygreaterthantheaveragerateatwhichtheintermediateunitsarrive,ascommonlyassumedintheoperationsmanagementliterature_HoppandSpearman2001_.

Knowledgeworkershavealotofpower,andyoudon'twanttoimposeth