《毕博上海银行咨询CreditRiskPresentationMay2002》由会员分享，可在线阅读，更多相关《毕博上海银行咨询CreditRiskPresentationMay2002（77页珍藏版）》请在金锄头文库上搜索。
1、ModelingCreditRiskinAsianContext:TheBuildingBlockApproachMay2002IntroductiontoCreditRiskModelingDataissuelHistoricalportfolioinformationoftennotavailablelApplicabilityofexternaldatais(highly)questionablelQualityofdataishighlyquestionableIntroduction:ChallengesinImplementingCreditRiskModelsinAsiaComp
2、arabilityofFinancialStatementslLackoftransparencylNopubliclytradeddebtlDisparityinaccountingstandardsUnpredictablerecoveryrateslTaxandregulatoryruleslBankruptcylawsunderdevelopedandincomparable注意注意CreditRiskModeling:ConceptualFrameworkandBuildingBlockApproachCreditRiskModeling:ConceptualFrameworkCre
3、ditriskmodelinghastobeenabledbyeffectivecreditriskinfrastructurewherecriticalhistoricalriskfactorsformodelingcanbestored.Toeffectivelytransformcreditriskmanagementofabank,itisimportanttointegratecreditriskmodelsandinfrastructureintoatransformedcreditriskprocess.CreditRiskDataStoreOperationalDataConv
4、ersionHistoricalDataConversionCreditRiskInfrastructureIdentificationofRiskCreditRiskModelModelDataBuildModelSelectionCreditRiskModelingQuantificationofRiskCreditRiskDecisionSupportLimit/RiskReportingCreditRiskProcessRe-designCreditRiskProcessIntegrationCreditRiskTrainingMonitor/ControlRisk注意注意 Model
5、ingCreditRisk:TheBuildingBlockApproachPhase 1 ComponentExternal Component Phase 2 ComponentPhase 3 ComponentCreditRiskProcessing13.ALCOorCapitalPlanningProcessDevelopTrainingProgramOutputManagement14.15.Phase3:ModelDeploymentandIntegration8.Setup/BuildCreditRiskModelingSystem11.ModelTesting12.BuildD
6、ataSet(SampleorPopulation)BuildDataSet(OutofSampleTest)10a.10b.ModelSelectionAcquireSoftware&Data(Optional)9.Phase2:ModelSelection,BuildingandTestingPhase1:CreditRiskInfrastructureDevelopmentCreditRiskManagementStrategy/PolicyDetermineCreditRiskModelingScope2.DetermineCreditRiskModelingObjectives1.D
7、ataManagement6.IdentifyCreditRiskFactorsAssessDataAvailability4.5.DesignSolutionArchitecture7.BusinessEnvironmentStudy3.方法方法Phase1:DesigningandImplementinganEffectiveCreditRiskInfrastructureProductCoverageConsumercreditCommercialcreditWholesalecreditClientTypeIndividualPrivateFirmPublicFirmInterbank
8、GovernmentCreditManagementProcessCreditOriginationCreditReviewProblemLoanManagementObjectivesRisk-adjustedpricingPortfolioriskdiversificationRiskcapitalplanningBIS2compliance(globalbestpractice)Cross-countryanalysisBlock1&2:DetermineCreditRiskModelingScopeExclusionsEmployeesDeceasedSpecificindustryM
9、odelingCreditRisk:Phase1-CreditRiskInfrastructureRealisticimplementationplan,typicallynotmorethan10creditriskmodelsforfirstattempt.ModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock3:BusinessEnvironmentStudyCreditRiskInfrastructureDesignCreditRiskManagementToolsDesignCreditRiskProcessIntegratio
10、nDesignEvaluateExistingCreditManagementSystemsandITstrategyDevelopITblue-printforCreditRiskSolutionArchitectureEvaluateExistingCreditRiskModelsDevelopDefinitionforExposure,ExposureAggregation(Related-Party),andCustomer,IndustryandProductSegmentationFeasibilityStudyofEmpiricalScorecardvs.SubjectiveJu
11、dgmentalRating-BasedApproachMapExistingCreditRiskManagementProcessesIdentifyKeyCreditControlsNo. Past Due 90d% 90d PD within IndustryNo. of Past Due 90d% of 90d PD within IndustryDistributionofNo.ofCorporateClientwithin90dPastDuebyIndustry:ModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock3:Bus
12、inessEnvironmentStudy(I)Wholesale & RetailManufacturingReal EstateFinancial ServicesConstructionProfessional ServicesSocial ServicesTelecommunication & ITUtilitiesFishery & FarmingFeasibility Cut-off注意注意DeterminationofCreditRiskModelingApproachbasedonBusinessCompositionStudyforCorporateLoanPortfolio
13、:ModelingApproachIndustryNoofModelsEmpirical Scorecard Model- Wholesale & Retail- Manufacturing- Real Estate- Financial Services4-5Subjective Judgmental Rating-Based Model- Construction- Professional Services- Social Services- Telecommunication and IT- Utilities- Fishery & Farming6ModelingCreditRisk
14、:Phase1-CreditRiskInfrastructureBlock3:BusinessEnvironmentStudy(I)ModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock3:BusinessEnvironmentStudy(II)18.9%approved loan were below cut-offExampleofEvaluationofanInternalSubjectiveScorecardModelofanAsianBankDevelopedforUnsecuredConsumerLoan:Excessives
15、ubjectiveinterventiontotheuseofscorecardwerefound17.5%above cut-off score were rejectedSubjective Cut-off注意注意ModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock3:BusinessEnvironmentStudy(II)ScoredistributionrevealedthatmodelwasnotabletodiscriminatebetweengoodandbadaccountsSubjective Cut-off注意注意M
16、odelingCreditRisk:Phase1-CreditRiskInfrastructureBlock3:BusinessEnvironmentStudy(III)ExampleofDefinitionofSmallandMediumSizedEnterprises(SMEs)wherebanksinternalclientinformationisnotavailable:Underlyingassumption:SMEsarenon-listedandnon-issuerofpublicdebtSegmentationapproach:Modellistedcompaniesonth
17、estockexchange,determine5%tailofsmallestlistedcompaniesascut-offcriteriaandmodelnon-listedcompanies,determine25%tailoflargenon-listedcompanies.Takelowestcut-offofthe2cumulativedistributionSegmentationcriteria:Grossrevenue,totalassetandtotalcapitalGrossRevenuecut-offatHK$10millionGrossRevenuecut-offa
18、tHK$20million注意注意ModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock3:BusinessEnvironmentStudy(III)Totalassetcut-offatHK$50millionTotalassetcut-offatHK$30millionTotalassetcut-offatHK$50millionTotalassetcut-offatHK$40millionCreditriskdataisakeydriverfordevelopingcreditriskmanagementcapability.Cre
19、ditriskdatacanbeeffectivelyusedfortestinginternalratingbasedmodelsaswellastobuildstatisticalmodelsforscoringcreditrisk.CRDSframeworkisacomprehensiveframeworkthatdefinethedatarequirementsfordevelopingcreditriskdatabasestosupportmodelbuildingandtesting.CRDSframeworkforcommercialloancomprisesofsevenbui
20、ldingblocks:CreditRiskDataStore(CRDS)FrameworkObligorDataFinancialDataDefaultStatusDataQualitativeDataFacilityDataInternalRatingDataRecoveryDataModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock4:IdentifyCreditRiskFactorsExampleofCreditRiskDataStore(CRDS)FrameworkforCorporateLoan:ModelingCredit
21、Risk:Phase1-CreditRiskInfrastructureBlock4:IdentifyCreditRiskFactorsCRDSframeworkforconsumerloancomprisesoffivebuildingblocks:CreditRiskDataStore(CRDS)FrameworkUp-to-datedemographicsofborrowersandguarantors=Updatefromhostsystemperiodically,e.g.monthlyLoanspecificinformationincludingaccountopeningdat
22、e,interestrate,collateralinformationetc.andperiodicaccountperformanceinformation,includingmonthlypayment,outstanding,delinquencystatusetc.Written-offaccountsdataandtheirrecoveryinformationshouldbestoredforfutureanalysis.DefaultinformationatcustomerlevelBorrower/GuarantorDataDefaultStatusDataLoanData
23、RecoveryDataScoreDataConsumerscoringdatahistoryModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock4:IdentifyCreditRiskFactorsQuantifiableRiskFactorsCurrent/QuickRatiosNPBT/AssetsNPBTYOYGrowthInterestCoverageSize(SalesorTA)DebtServiceCoverageInventories/COGSSalesYOYGrowthTL/TARE/AssetsNumberofPas
24、tdues/ExcessesIndustrySpecificRiskFactorse.g.CIDBgradeNon-QuantifiableRiskFactorsAudited/Qualified?Financialssubmittedwithin6monthsofFYEManagementExperienceSuccessionRiskFXRiskCountryRiskConcentrationRiskCommodityRiskSupplyRiskIndustryRiskImpliedAccesstoCapitalModelingCreditRisk:Phase1-CreditRiskInf
25、rastructureExampleofRiskFactorsforPrivateFirms:Block4:IdentifyCreditRiskFactors注意注意DemographicDataResidence(rentorown)YearsatcurrentaddressMaritalStatusOccupationEmploymentHistoryYearsoncurrentjobFinancialDataBorrowersearningsstabilitySizeoftheincomecushionrepresentedbydebtratios/savingsBorrowerssou
26、rcesofincomeOthercreditcards/loansExampleofRiskFactorsforConsumer:Block4:IdentifyCreditRiskFactorsCollateralDataSizeofaborrowersequityinahomeAgeofthemortgageGeographicallocationTypeofpropertySizeofpropertyLoantovalueratioModelingCreditRisk:Phase1-CreditRiskInfrastructureCommonIssues:Consistencyofris
27、kfactorscaptured(i.e.Samecalculationofratios/haircutsassumedetc)LegacysystemonlycapturesapprovedloanNohistory,rejectedcasesnotstoredSufficiencyofdefaultdataformodelbuildingCreditOriginationCreditReviewProblemLoanManagementAssessManualDocumentationAssessElectronicallyStoredInformationIdentifyInformat
28、ionGapsAgainstRiskFactorsIfAvailable:Recommenddataextractionandstoringapproach(warehousing)IfNOTAvailable:RecommendmanualdatacaptureorpurchaseIfCritical:Re-designCreditInformationManagementProcessReviewRiskFactorRequirementIfNon-Critical:RemovefromRiskFactorRequirementBlock5:AssessDataAvailabilityMo
29、delingCreditRisk:Phase1-CreditRiskInfrastructureModelingCreditRisk:Phase1-CreditRiskInfrastructureBlock6:DesignCreditRiskSolutionArchitectureLoanServicingSystemsMQSeriesTibcoMiddleware&ETLBack-endhostPresentationDecisionSupportCreditriskmodelingengineDataStoreandWarehouseCredit Risk Decision Support
30、 System (Score/Rating Application)Credit Risk DSS DatabasePhase2:ModelSelection,BuildingandTestingModelsavailableinthemarket:Statisticalapproachusingregressiontechniques(logit,probit,linear),neuralnetwork,decision-treetomodeltheprobabilityofdefault.Adependentvariableisexplainedbyasetofindependentvar
31、iables.Workbestwithlimiteddata.DiscriminantModelsRule-BasedorExpertModelsModelagainstastructuredprocessthatanexperiencedanalystusestoarriveatthecreditdecision.Modelischaracterizedbysetofdecisionrules.Assumptionsmadearenottestedandallocationofweightsfordecisionvariablesissubjective.Goodasfirstattempt
32、.MarketModelsModelscalibrateprobabilityofdefaultbasedonmarketprices,e.g.equitypricesorbondspreads.Modelsarelimitedtofirmswithsufficientliquidmarketpricesandwherefinancialmarketsareefficient.LimitedapplicationinAsia(ex-Japan).AgencyRating-BasedModelsModelsbasedonratingpublishedbyratingagenciesorinter
33、nalrating,incorporatinglossgivendefaultanddefaultprobabilityperratingscale,probabilitymatricesofratingmigrationandexpectedrecoveryperratingscale.Modelislimitedtoratedinstitutionsandisusedatportfolio-level.ModelingCreditRisk:Phase2-ModelSelection,Building&TestingBlock8:ModelSelectionRule-BasedorExper
34、tModelsModelIndividualPrivatePublicInterbankMarketModelsAgencyRating-BasedModelsDiscriminantModelsPartiallyApplicableFullyApplicableNotApplicableModelsavailableinmarket:ModelingCreditRisk:Phase2-ModelSelection,Building&TestingBlock8:ModelSelection注意注意IntegratedCreditRiskMeasurement:Block8:ModelSelec
35、tionModelingCreditRisk:Phase2-ModelSelection,Building&TestingCreditOriginationCreditReviewProblemLoanManagementProbabilityofDefault(POD)LossGivenDefault(LGD)ExpectedLoss=EADxLGDxPODExpectedLossRetailConsumerPrivateFirmPublicFirmWholesale注意注意Block8:ModelSelectionModelingCreditRisk:Phase2-ModelSelecti
36、on,Building&TestingKeyCreditRiskMeasurementDevelopmentProcess:Rules-BasedRatingInternalRating-BasedModelEmpiricalScorecardUnexpectedLossEstimationHistoricalDefault/RecoveryVolatilityorScenarioAnalysisCreditValue-at-Risk(CVaR)Internalrating-basedmodel,estimatingexpectedlossesandgeneratingcreditvalue-
37、at-riskarethreedistinctstagesofcreditriskmeasurementdevelopmentprocessHistoricalDefaultEventDatabaseHistoricalRecoveryEventDatabaseAccountLevelAnalysisProbabilitiesofDefaultEstimationLossGivenDefaultEstimationExpectedLossEstimationPortfolioLevelAnalysisModelingCreditRisk:Phase2-ModelSelection,Buildi
38、ng&TestingBlock9:AcquireSoftware&DataAndersenssystemselectionservices:Block10a:BuildDataSet(SampleorPopulation)Sampleorpopulationdependsondataavailability.Samplingmostlydoneinconsumer/retailportfolios.Assumptionsunderlyingselectionofsample.Proportionofdatainsampleagainstoutofsampledata.Block10b:Buil
39、dDataSet(OutofSampleTest)Outofsampledatasetisbuilttodetermineaccuracyofmodelbuilt.“Validation”ofmodels.Numberofdefaultdatainoutofsampledata.ModelingCreditRisk:Phase2-ModelSelection,Building&TestingBlock11:Setup/BuildCreditRiskModelingSystemBasedonacquiredsoftwareandmodelselected,Andersenwillsetupmod
40、elsforspecificportfolio/industry/geographicalsegments.ExampleofRiskFactorsetupisshownbelow:Block11:Setup/BuildCreditRiskModelingSystemExampleofmodelanalysisandperformanceisshownbelow:ModelingCreditRisk:Phase2-ModelSelection,Building&TestingBlock11:Setup/BuildCreditRiskModelingSystemExampleofsystemsc
41、orecardoutputisshownbelow:ModelingCreditRisk:Phase2-ModelSelection,Building&TestingModelingCreditRisk:Phase2-ModelSelection,Building&TestingLogicalTestApplicablefordiscriminantmodelonlyRelationshipforindependentvariablesmustbevalidated,e.g.higherleverageresultsinhigherprobabilityofdefaultNeedtodecid
42、etodropindependentvariable(s)Block12:ModelTestingStabilityTestAccuracyRatioTestStatisticalRobustnessTest,e.g.MaximumLikelihoodTestforLogitModelsModelRobustnessTestA)StabilityTestDescription:Aroutinetotestconsistencyoftheestimatedcoefficientsofthemodelconstructed.Typically,asampleformodelbuildisregre
43、ssed500times.Measurement:1)OrderTest-theorderofestimatedcoefficientsmustbeconsistentinall500run.Numberofrunwithinconsistentorderofestimatedcoefficientsarelogged.2)StandardErrorTest-eachestimatedcoefficientfromthe500runsismeasuredforthestandarderror.ModelingCreditRisk:Phase2-ModelSelection,Building&T
44、estingBlock12:ModelTestingB)AccuracyRatioTestDescription:Ideally,acreditriskmodelshouldreject100%ofbadcreditsfromsample(population)and0%ofgoodcreditsfromsample(population).Accuracyratioistotesttheperformanceonthemodelbasedoncut-offpointdetermined.Measurement: 1)FirstAccuracyRatioisdefinedasthereject
45、ionrateofbadcreditslessrejectionrateofgoodcreditsatagivenoptimizedcut-offpoint.2)SecondAccuracyRatioisappliedtoout-of-testsamplebasedonestimatedcoefficientsfrommodelbuild.DegradationoftheSecondAccuracyRatioismonitored.ModelingCreditRisk:Phase2-ModelSelection,Building&TestingBlock12:ModelTesting注意注意B
46、)AccuracyRatioTestAccuracyRatiodegradefrom73%to54%afterapplyingout-of-sampledata-set.ModelingCreditRisk:Phase2-ModelSelection,Building&TestingBlock12:ModelTesting?C)MaximumLikelihoodTestDescription:Maximumlikelihoodtestisatestagainstthesignificanceofhypothesisgivenadata-setusedformodelbuild.Measurem
47、ent: 1)TheLogofLikelihoodFunctionisdefinedasLog(P).GivenSisthenumberofsuccesses(Yt=1)observedinIobservations,thenforthelogitmodel,themaximumvalueoftheloglikelihoodfunctionunderthenullhypothesiscanbedefinedas:Atestofnullhypothesisthatalltheslopecoefficientsarezerocanbecarriedoutusingthelikelihoodrati
48、o(LR)test.LR test statistic is defined as 2Log(P) - Log(0). If themaximizedlikelihoodundernullhypothesis(H0),Log(0),ismuch smaller that the unrestricted maximized likelihood,Log(P). Therefore, a LR test statistic of more than zerowouldindicateanevidenceagainstthenullhypothesis.ModelingCreditRisk:Pha
49、se2-ModelSelection,Building&TestingBlock12:ModelTesting?Phase3:ModelDeploymentandProcessIntegrationCreditAcquisitionAccountMaintenanceCollectionsCreditRiskProcessingCreditRiskDSSWrite-offProductPlanning&PolicyIntegrationApplicationScoreAcquisitionPricingBehavioralScoreAttritionCross-sellFrauddetecti
50、onRecoveryScoreCollectionWorkoutModelingCreditRiskPhase3-Deployment&IntegrationBlock13:CreditRiskDecisionSupportSystem(CRDSS)ALCOIntegration-Volume,Rate,MixRAROCModelingCreditRiskPhase3-Deployment&IntegrationBlock13:CreditRiskDecisionSupportSystem(CRDSS)CRDSSstoresandintegratesthecreditpolicy,credit
51、modelsandbusinessrulesintobank-wideloanoriginationprocess.ModelingCreditRiskPhase3-Deployment&IntegrationBlock13:CreditRiskDecisionSupportSystem(CRDSS)AnexampleofcreditriskdecisionprocessenforcedbyCRDSS.ModelingCreditRiskPhase3-Deployment&IntegrationBlock14:DevelopTrainingProgramDevelopedofcreditris
52、kmodelingtrainingprogramiscrucialinthemodeldeploymentphase:Re-definitionofvariousfunctionsintheorganizationEssentialforbusinessunitsbuy-inandSeniorManagementendorsementTrainCreditOfficerstousethemodelTrainRiskOfficeronbeingadministratorsofthemodelTrainingisnecessarytofacilitatecreditofficersinmanagi
53、ngchange.ModelingCreditRiskPhase3-Deployment&IntegrationPortfolioConcentrationAssetQualityAnalysisDefaultHistoryProfitabilityAnalysisFacility&RecoveryHistoryModelPerformanceCompliance/ExceptionMonitoringLimitMonitoringBlock15:OutputManagementCRDSreportingframeworkcomprisesofeight(8)buildingblocks:Cr
54、editRiskDataStore(CRDS)ReportingFrameworkExposurebyRegion/Industry/SectorExpectedLossbyRatingGradesforRegion/Industry/Sector/TeamLeaderNumberofPastdues/ExcessesbyratinggradesforRegion/Industry/Sector/TeamLeaderTransitionMatricesStressTestchartsSampleRiskModelingReports:ModelingCreditRiskPhase3-Deplo
55、yment&IntegrationBlock15:OutputManagementMonitoringofExposure:33%ExposureConcentrationonCreditGrade732%ClientConcentrationonCreditGrade78%TotalExposureinSub-standardGrades注意注意MonitoringofPotentialLosses:$14mTotalExpectedLosses(%ofRiskCapital)$3.6mConcentrationofExpectedLossesonCreditGrade7CreditExce
56、ptionMonitoring:Reporttracksnumberofcompanieswithhistoricalpastdues/excessespercreditgradeSpecificattentiontoGrade6&7withatotalof34companieswith9timespastdues/excessesModelingCreditRiskPhase3-Deployment&IntegrationSummaryAstructuredandintegratedapproachtocreditriskmodelimplementationCreditriskinfras
57、tructure-weworkcloselywithITresourcesofclienttodeliveracomprehensivecreditriskinfrastructuresolution(includingmanualconversionofhistoricaldata)Creditmodelselection/build-recommendappropriatemodelgivendataconstraintandmodelingobjectives,implement/buildmodelandmanagemodeloutput/performanceCreditriskpr
58、ocessing-recommendappropriateapproachformodeldeployment,thisincludeanin-depthassessingofclientscreditcultureandITstrategyManageintegrationofcreditriskmodelingtobank-widecreditpolicy,creditapprovaldelegationauthoritiesandriskmeasurementTheKCIValueCaseStudy:ACreditRiskSystem(CreditRisk)DesignforaLarge
59、ChineseBankMay2002KeyDriversoftheProjectStrategicConsiderationsThe need to achieve the international standards in managing bank performance1nAnimportantcomponentforintegratedbankperformancemanagementisriskmanagement，itisalsoanareademandinghightechnicalcompetency.nToadoptglobalstandardformanagingecon
60、omiccapital.BankPolicyFormulationProcessBankOperationalControlAsset/LiabilityManagementProcessRiskAdjustedPerformanceReviewGuidelinesComplianceMonitoringBankStrategyFormulationProcessFinancialPerformanceReviewBudgetaryProcessCostAllocation(ABC)TransferPricingProfitabilityManagement注意注意StrategicConsi
61、derationsRegulatory CompliancenA required framework to determine financial soundness of Bank through effective quantification of risk.nProvide the necessary information to enable global investors to understand Banks risk appetite.Source:Bank of England Spring Quarterly Report,2001nThrough Basels req
62、uirement for internal credit risk rating, enable identification of creditworthy clients.nAchieve advanced approach to enable global investors to compare Bank with its peer banks.2The New Basel Capital Accord will affect Bank in the following ways:StrategicConsiderationsStrengthen Post-Merger Credit
63、Business Integration3nProvide risk measurement, analytical and monitoring tools to improve banks internal communication on risk, and reduce ambiguity and subjectivity in credit analysis. nProvide the necessary flexibility in managing credit policy, decision support and limit formulation process.nInt
64、egrate limit formulation and decision support across business units, products and channels.RiskManagementBusinessDevelopmentandManagementBankStrategyandPolicyFormulationProcessRiskdata,operationalfeedbackRiskpolicy,measurementtools,monitoringprocessEconomicCapitalAllocationRisk-AdjustedPerformanceMa
65、nagementUnexpectedLoss(Value-at-Risk)LoanPricingProfitabilityManagementIV.StrategicConsiderationsCompetitive Pressure4RiskDataCollectionandStorageI.InternalRiskScoringandRatingRisk-AdjustedLimitAllocationII.ProbabilitiesofDefaultandLossGivenDefaultExposure-at-DefaultExpectedLossIII.nMany internation
66、al banks have achieved Stage 4, and are in refinement stage. nTheir current activities, include refining their credit models through implementing a credit risk model for SMEs to complement large corporate clients. At the same time, they are improving their portfolio and banking institution risk meas
67、urement capabilities.CreditRiskProjectDeliverablesRiskexposurecoverage:Corporate (large and SME), sovereigns, consumer, project and equityfinancing,whilebankswillbeexcludedInstrumentcoverage:Termloans(securedandunsecured)SyndicatedloansGuaranteesissuedBillsandReceivablesNon-maturityloans(securedandu
68、nsecured),e.g.overdraft,lineofcreditSpecialisedloans(aspertheNewBasleAccorddefinition,i.e.project,realestate,objectandcommodityfinancing)SharemarginfinancingFXforwardsEquity investments (Direct or Indirect as per the New Basle Accorddefinition)Geographicalcoverage:Over300branchesinHKandChinaBusiness
69、ScopeSTARTHERETheCreditRiskProjectConceptualFramework：TheCreditRiskProjectModulesCreditRiskInfrastructureCredit Risk Data Store(Credit Risk/CBI) ElectronicDataConversionManualHistorical DataConversionRiskIdentificationData Capturing System (DCS) Credit Risk DecisionSupport System(DSS)CreditRiskProce
70、ssIntegrationCredit Risk Reporting using OLAPCreditRiskMonitoringCredit RiskLimit Management System(LMS)Internal Rating-Based ModelsExpectedLoss ModelsStress-TestingModels CreditRiskModelingRiskMeasurementCreditRAROCUnexpectedLoss ModelsFundamental New Basel Capital Accord RequirementComprehensiveri
71、skmeasurementtoolsthatcomplywiththeNewBaselCapitalAccordandinternationalbestpracticenReflectKCIsunderstandingofAsianbanksnIntegratewithbankscreditbusinessandriskmonitoringprocessRisk IdentificationnBanksBusinessCompositionStudyCreditindustrysectorandproductcompositionLoanvolumeandriskmodelingfeasibi
72、litystudyRelated-partydefinitionAssessmentofbanksexistingmodelsCreditriskfactordefinitionnDataaggregationapproach,includingelectronicandmanualdataconversionnApproachforone-offhistoricaldataconversionnApproachandstandardforelectronicdataextraction,transformationandloading(ETL)nLogicaldatamodelforstor
73、ingcreditriskfactorsCreditRiskProjectDeliverablesRisk MeasurementnInternalrating-basedmodelDefinitionofnon-performingcustomerObligorriskratingEmpiricallargecorporatecreditriskmodelEmpiricalsmallandmediumenterprisecreditriskmodelEmpiricalconsumercreditriskmodelProjectfinancecreditriskmodelbasedoncash
74、flowprojectionSubjectiveconstraint-basedfacilityriskratingmodelEmpiricalcountryriskratingmodelInternalsubjectiveportfolioriskratingmodelIntegratinginternalriskratingapproachInternalriskratingmethodology,includingratingoverrideandadjustmentCreditriskmodeltestingmethodologyIntegrationofinternalrating-
75、basedapproachwithexistingregulatoryratingapproachnExpectedlossmodelExposure-at-defaultdefinitionbyproductProbabilitiesofdefaultandlossgivendefaultestimationapproachExpectedlosscalculationapproachCreditRiskProjectDeliverablesRisk MeasurementnInternalrating-basedmodelstress-testingapproachnUnexpectedl
76、ossestimationEvaluationofCreditMetricandCredit+approachRecommendedCredit+implementationapproachnCreditRAROCmethodologyIntegratewiththebanksprofitabilitymanagementapproachDesigneddatacollectionandcreditRAROCestimationapproachCreditRiskProjectDeliverablesnCreditRiskDecisionSupportSystem(CRDSS)Internal
77、creditscoringandratingcalculationapproachExpectedlosscalculationapproachCreditRAROCcalculationapproachLoanpricingapproachCreditpolicycomplianceapproachnCreditLimitManagementSystem(LMS)Related-partyandgroupdefinitionGroupandconsumercustomerlimitcalculationapproachCreditriskconcentration,portfolioandp
78、olicylimitLimitimplementationandmonitoringapproachRisk MonitoringnCreditriskmonitoringandmanagementqueryComprehensivecreditriskreportAutomateselectionregulatoryandHeadOfficereportSelectivecreditbusinessreportCreditdataquerynCreditprocessintegrationDocumentedexistingcreditprocess,includingidentificat
79、ionofcreditcontrolpointandownerIdentifiedCreditRiskintegrationpointwithcreditprocessCreditRiskProjectDeliverablesnCreditRisksolutionarchitectureframeworknCreditRiskbusinessrequirementnCreditRisksystemselectionCreditRiskrequestforproposal(RFP)AssessmentofvendorsproposedsolutionsnCreditRiskfunctionals
80、pecificationnCreditRiskprototypesolutionsnCreditRiskbusinessimplementationsupportingplanBanksresourceandknowledgetransferplanOrganizationalstructureandpolicyconsiderationsManualdataconversiontasklistnCreditRiskPhase2implementationprojectplanandcostestimationCredit Risk Overall DesignCreditRiskProjec
81、tDeliverablesCreditRiskSolutionPrototypeCreditRiskInfrastructureCredit Risk Data Store(Credit Risk/CBI) ElectronicDataConversionManualHistorical DataConversionRiskIdentification CRPMS Data Capturing System (DCS) CRPMS Credit Risk DecisionSupport System(DSS)CreditRiskProcessIntegrationCredit Risk Rep
82、orting using OLAPCreditRiskMonitoring CRPMS Credit RiskLimit Management System(LMS)Internal Rating-Based ModelsExpectedLoss ModelsStress-TestingModels CreditRiskModelingRiskMeasurementCreditRAROCUnexpectedLoss ModelsFundamental New Basel Capital Accord RequirementCredit Risk prototype demonstration
83、process:CreditRiskPrototypeSolutions2 23 34 45 56 67 78 89 91 1CreditRiskPrototypeSolutionsCreditRiskModule1：DataCaptureSystem(DCS)CreditRiskDataCreditBusinessDataDocumentComplianceBlackListClientProfitabilityDataDepositDataObligorDataFacilityDataDefaultStatusDataFinancialDataRecoveryDataQualitative
84、DataRatingAndLimitDataModelParameterCreditRiskPrototypeSolutionsCreditRiskModule1：DataCaptureSystem(DCS)(contd)CaptureofObligorDataSampleCreditRiskModule1：DataCaptureSystem(DCS)(contd)CreditRiskPrototypeSolutionsSampleDCSderivescashflowstatementfromhistoricalbalancesheetandincomestatementandabletoge
85、nerateratioanalysis.CreditRiskPrototypeSolutionsCreditRiskModule2：CreditRiskDataStoreSampleIntegratedandcomprehensivecreditriskdatabasetosupportqueryandresearchanalysis.CreditRiskPrototypeSolutionsCreditRiskModule3：InternalRatingBasedModelsEmpiricalmodeldevelopmentprocess：SelectTransformMineAnalysis
86、SelectDataTransformDataAssessandDevelopDataPatternDevelopDataAnalysisModelCreditRiskPrototypeSolutionsCreditRiskModule3：InternalRating-BasedModelsSampleEmpiricalmodeldevelopmentandtesting.CreditRiskPrototypeSolutionsCreditRiskModule3：InternalRating-BasedModelsSampleTheprojectfinancecashflowsimulatio
87、nmodel.CreditRiskPrototypeSolutionsCreditRiskModule4：UnexpectedLossModelUnexpectedlossportfolioeconomiccapitalcalculation.SampleCredit loss distributionUnexpected loss at 99% confidence intervalExceptional loss beyond 99% confidence intervalCredit VaR at 99% confidence intervalPortfolio average loss
88、ReturnPortfolioExposurePortfolioAvg.LossPortfolioLossProb.CreditRiskPrototypeSolutionsCreditRiskModule5：DecisionSupportSystem(DSS)CreditRiskPrototypeSolutionsCreditRiskModule6：ExpectedLossCalculationCreditRiskPrototypeSolutionsCreditRiskModule7：CreditRAROCandLoanPricingModelCreditRiskPrototypeSoluti
89、onsCreditRiskModule8：LimitManagementSystem(LMS)CreditRiskPrototypeSolutionsCreditRiskModule9：CreditRiskReportingandQuery(OLAP)ConcentrationDefaultHistoryAssetQualityFacilityandRecoveryHistoryPolicyComplianceandExceptionLimitMonitoringRisk/ReturnAnalysisModelPerformanceandTestingnEnhanced Credit Risk
90、 ReportsnAutomate selective credit business, Head Office and regulatory reports.CreditRiskPrototypeSolutionsCreditRiskModule9：CreditRiskReportingandQuery(OLAP)MajpjMVcyzj21HLfrvy96dv02lPPfYgxUS7IYmZkyEmZ0kGeYZS3bpLCkYH1lt4EK7CxmUX3ijoYSOer7ZuaVWYgz4EpZrUirVpMzzvNtf1XZw5oswSXOtFaejnOcmfE1lZgnN1RSXg8w
91、LCG8CVQ3XPJMvodPFWcpiYJgZazNSEPNIaklYSu7qSd1UpaxmZDlpN9zW7kljfsLCLi26Yv109ffbnDH8LbUN1G6ACURQ39eG12KHL9tXsZ1jzgoCK8g1kuNOh5eFvcmVT5ZYVQt9zk3rp3qLnf02FovEXxVRxjCcFRNppiJljNiOuk6fONnyX7fyGg7sXZ49BmCN5oy9VesHpKzdjTKwjrkCEQCFDehVmGax3lrOEbw63VscA3YSijtUKoCyiLzAlVRp7l4QgPNHxvJFFDyjUVN3oHlMah0XBd4uTbkfPIh
92、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
