消费信贷风险管理[文献翻译]VIP免费

ManagingConsumerCreditRiskPeterBurnsAnneStanleySeptember2001AbstractOnJuly31,2001,thePaymentCardsCenteroftheFederalReserveBankofPhiladelphiahostedaWorkshopthatexaminedcurrentcreditriskmanagementpracticesintheconsumercreditindustry.ThesessionwasledbyJeffreyBower,seniormanagerinKPMGConsulting’sfinancialservicespractice.Bowerdiscussed"bestpractices"inthecreditriskmanagementfield,includingcreditscoring,lossforecasting,andportfoliomanagement.Inaddition,heprovidedanoverviewofdevelopingnewmethodologiesusedbytoday'sriskmanagementprofessionalsinunderwritingconsumerrisk.ThispapersummarizeskeyelementsofBower'spresentation.Keywords:consumercredit,creditscoring,portfoliomanagementPORTFOLIOMANAGEMENTANDANALYTICSInBower’sview,consumercreditriskmustbeunderstoodintermsofaportfoliomanagementstrategythatbalancescapitalpreservationwithcapitaloptimization,thatis,“...acontinuousprocessofidentifyingandcapitalizinguponappropriateopportunitieswhileavoidinginappropriateexposureinsuchawayastomaximizethevalueofenterprise.”Capturingdataacrossallstepsinthecustomerrelationshipandintegratinginformationmanagementarethekeystoeffectiveportfoliomanagement.Whilethisisafairlystraightforwardprescription,executingitisoftenbeyondthescopeofmanylenders,withthecreditcardcompaniesgenerallyinthevanguard.Often,theprocessstepsaremanagedonseparatelegacysystems,whichcomplicateeffortstointegrateinformation.KPMGconsultantsfindthatmanyfirmstypicallypurgespecificfilesbeforetheinformationisextractedandcombinedwithotherdatatoprovideeffectiveportfoliomanagement.Thelossofapplicationdata,forexample,wouldmeanthatcriticalscore-cardanddemographicinformationwouldnotbeavailabletomodelbehaviorindefinedcustomerorrisksegments.BESTPRACTICESINCREDIT-RISK-MANAGEMENTCredit-risk-managementpracticesvaryconsiderablyamongfirmsandbetweensegmentsoftheconsumerlendingindustry.Toillustratethevariability,Bowerdescribedtherangeofmanagementpracticesin:Creditdecision-makingcredit-scoringlossforecastingportfoliomanagement.Onthefrontendofthecreditprocess,industryleadersareinvestinginanalyticstoimprovethecreditdecision-makingprocess.Buildingonexperiencewithcreditscoringtechnologies,theseleadersareemployingexpertsystemsthatcanadopttochangesintheeconomyorwithinspecificcustomersegments.Theuseofcredit-scoringvariesfromthosethatareusingonlycreditbureaudata,tothosethatblendbureaudatawithotherinformationbasedonthefirm’sownexperience,tothemostadvancedapplicationsusingadaptivealgorithms.Thesemodels,usedbysomeoftheleadingcreditcardissuers,areupdatedregularlytoreflectchangingcharacteristicsoftheapplicantpopulation.Asignificantchallengeforeventhemostsophisticatedlendersishowtomodelprobableperformancewhendealingwithnewcustomersegments.Lossforecastingtechniqueshaveadvancedconsiderablyfromtheirearlyrelianceonhistoricaldelinquencyratesandcharge-offtrendanalyses.Delinquencyflowmodelsandsegmentedvintageanalysesarenowcommonlyusedtorecognizeportfoliodynamicsandbehaviorpatternsbasedonpoolswithcommoncharacteristics.Thecreditcardindustryhasperhapsgonethefurthestwithitsuseofmassivesegmentationprofiles,withthemoreadvancedissuerscomplementingtheseprofileswithregionaleconomicdataandotheranalyticaldynamics.Overallportfoliomanagementemploysallofthesetechniqueswithmostfirmstrackingcurrentvs.historicalperformanceandestablishingconcentrationlimitsforparticularrisksegments.Somelende...