商业银行资信评估及实证分析

华中科技大学硕士学位论文商业银行资信评估及实证分析姓名：刘玥申请学位级别：硕士专业：金融学指导教师：田新时20070605I2006QIIFisher()FisherFisherIIIAbstractCorporatecreditratingisakindofeconomicactivitytoevaluatethecapabilityofrepayingandtherepayingwillshowtheevaluatingresultwithsimplecodewithrelevantfortheinvestor,supervisioninstitutionandotherrelativesides.Commercialbankcreditratingisbasedontheanalyzingalltheriskfacedbythebank,usingsomemethodtomeasuretheprobabilityofdefaultandcapabilityofrepayingandusingasimplecodetorepresentforit.Commercialbankisimportantforthewholefinanceindustry,andithasahugeinfluenceonthedomesticeconomicdevelopment.Commercialbankcreditratingprovidesriskinformationandsuggestionsforthedepositors,investorsandsoon.Atthesometime,commercialbankcreditratingishelpfulforthebanksupervisioninstitutiontosuperviseandcontroltherisk.Itisagoodwaytokeepthecitizenhaveconfidenceonthebanks,sotheeconomicwilldevelopinasaveway.Tothecommercialbankitself,thecreditratingwillhelpittosetagoodimagetothepublic,enlargethemarketandreducethecost.Tillnow,theattentionpayedtothecreditratingisnotenough,especiallythecommercialbankcreditratingandtheresearchinthisareaislittle.Asthefinancialmarketbecomemoreandmoreinternational,thecommercialbankcreditratingwillbecomemoreandmoreimportant.Anditisessentialtofindagoodwaytoevaluatethecreditratingofourcommercialbank.Thepaperanalyzedthecommercialbankcreditratingfrombothqualityandquantityaspects.Fromthequalityaspect,weanalyzethecommercialbankcreditfromsixareas:economicenvironment,industryenvironment,financialsupervisionenvironment,risk,managementandgovernmentsupport.Inthequantityaspect,webuildamodelusingfinancialdatatomeasurethecreditrate.Thearticleusedthestatisticsmethod:thefactoranalysisisusedtoextractthecommunalcomponents;thentheclusteranalysisisusedtoseparatethesamplebanksintothreekinds;ontheclassifyingthediscriminantanalysisisusedtoattaintheFisherfunctions;Themainpurposeofthismethodistosupplyasimpleandfeasiblemethodtoevaluatethebank'screditrating.Astheresultofthearticle,wegottwofunctionstoevaluatethecommercialbank’screditrating.Oneisusingtheresultoffactoranalysis,whichismoreaccuratebutcomplextocompute.Fromthismethod,wecangetthecreditscoreforeachbank.TheotheroneistheFisherfunction.Weusethismethodtoclassifythebanksintoseveralranks.Keyword:commercialbankcreditratingqualityanalysisquantityanalysis□_____□“√”111.120071——21.21.2.119661954-19647914“/”196831946-196533225ZZ1977(ZETA)1980Logit492%90J.P.Morgan(1997)CreditMetricsKMVKMVMcKinseyCreditPortfolioViewCreditRisk+KPMG1J.P.MorganCreditMetrics2RNArrow(1953)HarrisonandKreps(1979)HarrisonandPliska(1981)Kreps(1982)3BSMstructuremodelKMVreduced-formmodelCreditRisk+41.2.22005(2002)CreditMetrics}CreditMonitorTMCredtRisk+,CreditPortfolioViewTM,LoanAnalysisSystem1.31.3.15QFisher()1.3.2(1)(2)(3)2006fisher622.172.12.212382.32.3.12.3.2129,,WTO2.3.3102.3.4(1)(2)a)b)c)VaR)(3)(4)(5)112.3.542.3.61233.13.1.119663.1.2Fisher1936LogitProbit,Logit·(EdwardLAltman)1968SZ-score“Ze-to”133.1.3QuinlanHuntCLS(ConceptLearningSystem);GreeneSmith(GA,GenerationAlgorithm):RomaniukHallFuzzynet143.1.4RAROCRAROC2070RAROC()Sharpe(Risk-adjustedreturnoncapital):RAROC=()/()KPMG3.215QFisher()3.33.3.1“”1931Thurstone20KarlPearsonCharlesSpearmen:np01.xY1Y2……YmFlF2Fm(mP):(1)x=(x1,x2,……,xp)′E(x)=0cov(x)=EER;(2)F=(F1F2Fm)′mpE(F)=0cov(F)=IF;(3)ε=(ε1ε2εp)FE(ε)=0,εεxi=aijF1+aijF2+...+aijFmj+εi:i=1,2,…p;j=l,2,……,m;(4):F=a1F1+a2F2+...+amFm:al,a2,……,am;F1,F2,……,Fm;F163.3.2()Q():R-()K(K-MeansCluster):(ClusterCenter)(ClusterAnalysis)“”()()()()()()()3.3.3(DiscriminantAnalysis,DA)(R.A.Fisher)193617(discriminantfunction)“”:kX1,X2,……,XkF1(x)F2(x)…Fk(x)Fi(x)mx:Y=alXl+a2X2+……+anXn:Y()Xl……Xna1……anSPSSmm:“”“”3.41820063.5133.5.14%8%CR1CR2CR1=/CR2=/193.5.2=/=/3.5.3=/=/=–/=/3.5.420=—/=3.5.5=/=/3.63.6.11313SPSS213.1InitialExtraction1.000.866/1.000.9291.000.8451.000.8961.000.8581.000.9041.000.6931.000.8211.000.7591.000.7811.000.3681.000.259()90.4973223.2ComponentInitialEigenvaluesExtractionSumsofSquaredLoadingsRotationSumsofSquaredLoadingsTotal%ofVarianceCumulative%Total%ofVarianceCumulative%Total%ofVarianceCumulative%16.75261.38561.3856.75261.38561.3854.60341.84941.84922.16119.64781.0322.16119.64781.0322.73324.84266.69131.0419.46490.4971.0419.46490.4972.61923.80690.4974.4273.88394.3805.3152.86597.2466.1721.56398.8087.093.84999.6578.031.27899.9359.007.065100.000103.118E-162.835E-15100.000112.212E-162.011E-15100.0003.3Component123.718.225.461.197.863.327.517.019.825-.450-.762-.349.900-.324.154.749.479.354-.084.925-.202/.802.097.556-.876-.290-.259-.871-.228-.138.183.144.955F12322F10.8F1xi41.849;F2F2xi24.842:F3F1xi23.8063.4F1F2F341.84924.84223.80690.4970.460.280.261F=0.46F1+0.28F2+0.26F3F:F1,F2,F330.46,0.28,0.26F1F2F333F1=3.84F2=5.95F3=8.41,F=0.46*3.84+0.28*5.95+0.26*8.41=5.62.3.6.2F1F2F3QWardMethod()SPSS10243.1DendrogramusingWardMethodRescaledDistanceClusterCombineCASE0510152025LabelNum+---------+---------+---------+---------+---------+6ØÞ10Øà5ØÚØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØØÞ9ØÝÙ3ØÞÙ8ØÚØÞÙ2ØÝßØÞÙ7ØØØÝßØØØØØØØØØØØØØØØØØÞÙ1ØØØØØÝßØØØØØØØØØØ